Compare commits

...
Author SHA1 Message Date
Jason WoltjeandClaude Fable 5 90b3ab8d53 test(mutator-gate): keep launch ledger out of the shipped framework tree
ci/woodpecker/pr/ci Pipeline failed
runRuntimeLaunchEntry set MOSAIC_HOME to the shipped framework root, so
launch-runtime.py appended its launch ledger to
framework/fleet/run/sessions/events.ndjson — polluting the tree that
manifest.spec.ts walks and failing its completeness check in CI.

Point MOSAIC_HOME at the per-entry temp root instead. Nothing in the
launch chain resolves tools via MOSAIC_HOME (entry scripts resolve via
SCRIPT_DIR); the ledger is its only consumer here.

Co-Authored-By: Claude Fable 5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01Dtdjx4Gxude9fwyLezCrhh
2026-08-11 17:35:10 -05:00
mos-dt-0andClaude Fable 5 f6f846b92a test(skill): install linker asserts harness-home link topology
ci/woodpecker/pr/ci Pipeline failed
The two install-linker-compatibility tests still asserted the pre-isolation
behavior (mosaic skill links planted in $HOME/.claude/skills). This branch
deliberately moved the link farm into the mosaic-owned harness homes
($MOSAIC_HOME/.claude/skills) and demoted the base-install dirs to
cleanup-only legacy targets, so the tests now assert the new topology:
the skill links appear under the harness home, foreign links in the legacy
dir are preserved, and no new mosaic link is planted in the base install.

Co-Authored-By: Claude Fable 5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01Dtdjx4Gxude9fwyLezCrhh
2026-08-11 17:11:11 -05:00
Jason Woltjeandmos-dt-0 accf2f70ad feat(mosaic): mechanically authorize lease promotion 2026-08-11 17:04:22 -05:00
52f8278bc2 test(lease): #1124 regression uses node pane command (real field topology per scooby)
The mosaic wrapper makes pane_current_command=node (RUNTIME_ACCEPTABLE_COMMANDS.claude=['claude','node']); the walk matters precisely in that no-shell-wrapper case. Match reality.

Co-Authored-By: Claude Opus 4.8 <[email protected]>
Claude-Session: https://claude.ai/code/session_013SAYFkRhQfhguY7AHfiUC8
2026-08-11 17:04:22 -05:00
fff273dfbd fix(lease): resolve lease session id from the claude child, not the tmux pane pid (#1124)
The launcher runs the runtime as a spawnSync CHILD of node(mosaic) (deliberate,
per launch.ts:99 — parent survives to propagate signals), so
MOSAIC_LEASE_SESSION_ID lives on the claude child, not the pane's root pid. The
transport read only pane.pid's /proc/environ and returned RESOLVE_FAILED for
every real 'mosaic claude' seat. Now BFS the pane's process subtree (bounded,
injectable children-reader) and read the first descendant that carries a valid
lease id; fail-closed if none. Unit tests now exercise the real walk (pane=node
without lease -> child=claude with lease) rather than mocking the resolution.

Found by scooby greenfield E2E on fomo-lin with proc-level evidence.

Co-Authored-By: Claude Opus 4.8 <[email protected]>
Claude-Session: https://claude.ai/code/session_013SAYFkRhQfhguY7AHfiUC8
2026-08-11 17:04:22 -05:00
Jason Woltjeandmos-dt-0 83266661e4 test(lease): align mutator carve-out acceptance 2026-08-11 17:04:22 -05:00
Jason Woltjeandmos-dt-0 15b56384e0 fix(mosaic): bound promotion transport delivery 2026-08-11 17:04:22 -05:00
Jason Woltjeandmos-dt-0 d8a28efb3e feat(mosaic): add correlated lease promotion CLI 2026-08-11 17:04:22 -05:00
Jason Woltjeandmos-dt-0 73a95b4dde feat(lease): add single-turn Claude promotion trigger 2026-08-11 17:04:22 -05:00
0e0212e3f8 fix(lease): raise lease TTL 300s -> 3600s
MAX_LEASE_TTL_SECONDS (daemon cap+default) and DEFAULT_TTL_SECONDS
(lease_promote client) both move to 3600. The 5-minute TTL made
gated-by-default sessions unusable (re-promotion mid-task); 1 hour
matches a working session. Full test:framework-shell RC=0.

Co-Authored-By: Claude Fable 5 <[email protected]>
Claude-Session: https://claude.ai/code/session_013SAYFkRhQfhguY7AHfiUC8
2026-08-11 17:04:22 -05:00
Jason Woltjeandmos-dt-0 4c894c7fc9 fix(lease): ignore benign observer idle replies 2026-08-11 17:04:22 -05:00
Jason Woltjeandmos-dt-0 2b2f8aac7c fix(lease): constrain read-only tool carve-outs 2026-08-11 17:04:22 -05:00
Jason Woltjeandmos-dt-0 e4bafabe18 test(lease): assert pi carve-out capability 2026-08-11 17:04:22 -05:00
Jason Woltjeandmos-dt-0 58a6def568 test(lease): distinguish pi probe timeouts 2026-08-11 17:04:22 -05:00
Jason Woltjeandmos-dt-0 960b14d138 test(lease): enforce read-only tool invariant 2026-08-11 17:04:22 -05:00
Jason Woltjeandmos-dt-0 187f20bfb7 docs: measure pi tool registry 2026-08-11 17:04:16 -05:00
8fdc4b7180 test: close W-0R review findings — assert the omission notice, skip chmod simulations under root
The independent W-0R review of 3592b92e passed but left two PLAUSIBLE
findings: the stderr notice for a legitimately-omitted operator source was
claimed and never asserted (a silent omission is the original defect in
miniature), and the chmod 0o000 unreadable simulations fail spuriously when
euid==0 (CAP_DAC_OVERRIDE). Falsifier for the new assertion: deleting the
notice block turns the suite red (failures=3); restoring returns green.

Co-Authored-By: Claude Fable 5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01EHYXhcCQsL3J1Lnm7EraGq
2026-08-11 17:04:16 -05:00
718dbc9c0b fix(lease): refuse an incomplete law binding instead of silently shrinking it
build_construction skipped any normative source it could not read
(`except OSError: continue`) and promoted whatever remained. That is not a
degraded binding, it is a forged smaller one: the broker recomputes h_source /
h_payload from the fragments it is SENT (daemon.py:602-616), so an omitted
fragment is internally consistent and PAYLOAD_BINDING_MISMATCH cannot fire. A
partial law promotes exactly like a complete one and nothing downstream can tell
the difference.

Measured before this change, against a seeded home: with only USER.md readable,
the client produced a one-fragment construction with promotion=True. Removing
CONSTITUTION.md, STANDARDS.md or the runtime contract likewise promoted.

The classification mirrors the framework's own file ownership rather than
inventing one:

  * CONSTITUTION.md / AGENTS.md / STANDARDS.md are framework-owned and
    reconciled every upgrade (install.sh FRAMEWORK_OWNED,
    config/file-adapter.ts FRAMEWORK_OWNED_FILES), as is the per-runtime
    RUNTIME.md. Absent => IncompleteBinding. A deployment missing one is broken,
    not minimal.
  * SOUL.md / USER.md are deliberately not seeded by install.sh ("generated by
    `mosaic init`") and TOOLS.md is seeded on first install only, so their
    absence is legitimate. It is reported on stderr, never silent.

Unreadable is handled separately from absent for EVERY source, optional ones
included: a file that will not open is not a file that was never configured, and
collapsing the two is what let a permission change quietly shrink the law.

Also corrects this module's own docstring, which asserted that a VERIFIED lease
means "this agent is running THIS law". It does not. Both sides of the broker's
comparison originate in this client, so it detects corruption in transit and
nothing else. That overstatement is where the belief spread from; the stronger
claim needs the broker re-reading on-disk sources against a manifest the agent
cannot rewrite.

Test: promotion_binding_unittest.py, enumerated in test:framework-shell (the
enumeration guard's population is *test*.sh and does not cover Python, so an
unenumerated test here would simply never run). Falsifier executed: defeating the
guard while leaving the module API intact turns the suite red (12 failures);
restoring it returns green.

Co-Authored-By: Claude Opus 5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01EHYXhcCQsL3J1Lnm7EraGq
2026-08-11 17:04:16 -05:00
Jason Woltjeandmos-dt-0 0bca3ebd76 Revert "feat(pi): wire lazy lease promotion into the mutator gate"
This reverts 939f2e04. Keeping the revert rather than dropping the commit,
because the failed attempt is the most useful record on this branch.

The wiring worked mechanically — verified with a live model on sb-it-1-dt: the
receipt was emitted verbatim as a whole message, and the broker token was minted
AND consumed, so observe_receipt and promote_lease both succeeded and the lease
reached VERIFIED.

It failed as a DESIGN, for reasons that are properties of the protocol rather
than of this wiring:

  * It puts control-plane traffic in the user-facing conversation channel. An
    operator asking "what model are you?" received a receipt string instead of an
    answer — the model tried a tool, was blocked, complied with the receipt
    instruction, and in one-shot mode that text turn BECAME the reply. Observed
    twice, non-deterministically.
  * The lease TTL is hard-capped at 300s (MAX_LEASE_TTL_SECONDS; ttl_seconds >
    cap raises INVALID_LEASE_TTL). Measured: allowed at T+0, LEASE_EXPIRED at
    T+310. So the visible cost recurs every five minutes of mutator activity.
  * Model compliance is not guaranteed — one run retried the command instead of
    emitting the receipt.

Any model emission is user-visible, so this is not fixable by better wiring; it
needs a design answer about how promotion is triggered and paid for. That is
under adversarial review (docs/scratchpads/lease-remediation/07-liveness-design-brief.md
in the operator's repo). Promotion triggering will return on its own branch once
that lands.

What remains here is independently sound and unblocked: harness-home isolation,
the immutable launch record, the skills relocation, the promotion client itself
(steps 1/4/5), and the #1087 prefix guard.
2026-08-11 17:03:58 -05:00
Jason Woltjeandmos-dt-0 7dc99470bd fix(sync-skills): guard the pre-existing prune against an empty prefix (#1087)
prune_stale_links_in_target compared "$resolved" == "$canonical_real/"* while
length-checking only $resolved. If $canonical_real were ever empty the pattern
collapses to == "/"* and matches every absolute path.

The failure is precisely inverted, which is what makes it worth fixing rather
than noting: is_mosaic_skill_name already `continue`s for names that ARE current
mosaic skills, so an empty prefix would delete exactly the FOREIGN symlinks in
every target directory and preserve the mosaic ones. On this host that is 4 base
installs, including codex's own .system entry.

Reported by mos-claude as #1087 after I introduced the same guard in the new
legacy-cleanup path in the previous commit and walked past this instance thirty
lines away. Same defect class, same file, one function apart.

$canonical_real is populated by readlink -f after a mkdir -p, so an empty value
requires readlink to fail — unlikely, but the consequence is deleting operator
symlinks across every harness, which is not a risk worth carrying for one test.
2026-08-11 17:03:58 -05:00
Jason Woltjeandmos-dt-0 c9913fa4b1 feat(pi): wire lazy lease promotion into the mutator gate
Completes the promotion path: the client landed in the previous commit, but
nothing drove step 2 — the model emitting the receipt. This wires it.

LAZY, not at session start. Promotion costs an entire model turn, because the
receipt must be the whole message (hmac.compare_digest, "not a transcript
substring"). Minting at session start would collide with the Constitution's
first-response mode declaration — the two cannot share a message, so requiring
both would be unsatisfiable. Deferring to the first DENIED MUTATOR means the
mode declaration happens first and the receipt gets its own later turn, so no
governance change is needed. A read-only session never pays for promotion at all.

Mechanism: on a MUTATOR_UNVERIFIED denial the tool_call hook mints a challenge
and returns the receipt in the block `reason`, which pi feeds back to the model
as the tool result — the existing injection path already used by
lease-lifecycle.ts. The model emits the receipt as its next message, message_end
ships it to the observer, and the extension then calls observe_receipt +
promote_lease.

Only MUTATOR_UNVERIFIED triggers minting. Other denials (GATE_UNAVAILABLE,
STALE_GENERATION, LEASE_EXPIRED, ANCESTRY_MISMATCH) describe conditions a
receipt cannot fix, and begin_verification revokes before it mints, so minting
there would thrash the broker.

Completion is gated on an EXACT text match against the minted receipt. This is
load-bearing, not defensive: message_end also fires for the message that
CONTAINED the blocked tool call — one turn BEFORE the model answers. An earlier
version completed there, so observe_receipt compared against the wrong text,
failed, and burned the challenge before the model ever emitted it. Matching the
text mirrors the broker's own compare_digest semantics and waits for the right
turn. Confirmed by instrumenting message_end and watching it fire with
pending=yes one message too early.

It never posts the receipt itself. receipt-observer-client.py accepts any
string, so self-posting would satisfy the broker while proving nothing — the
whole point is that a live model echoes a challenge it was given.

Bounded by MAX_PROMOTION_ATTEMPTS: model compliance is not guaranteed (observed
a run where the model retried the command instead of emitting the receipt), so
a non-complying model degrades to today's behaviour — denied mutators — rather
than looping.

Verified with a live model on sb-it-1-dt: receipt emitted verbatim as a whole
message, and the broker token was minted AND consumed, i.e. observe_receipt and
promote_lease both succeeded and the lease reached VERIFIED.

Known limitation: under `pi -p`, the receipt is a text-only turn, which ends the
one-shot loop — so promotion completes but the blocked tool is not retried in
that same invocation. Interactive and durable fleet sessions continue and retry
normally.
2026-08-11 17:03:58 -05:00
Jason Woltjeandmos-dt-0 226a9d19d9 feat(lease-broker): add the missing promotion client
The enforcement half of the lease broker ships and denies; the promotion half
has no production caller anywhere in the package. Verified across 0.0.48, 0.0.49
and 0.0.50-next.2207: begin_verification / observe_receipt / promote_lease are
invoked only by broker-test-client.ts, the acceptance spec, unit tests, and two
probes under docs/.

Consequence: no lease on any host can reach VERIFIED, so mutator-gate denies
every mutator with MUTATOR_UNVERIFIED via a gate that nothing shipped can
satisfy. Runtimes that enforce the gate in-process (pi, via mosaic-extension's
tool_call hook) are bricked for mutators; runtimes whose gate is wired through a
settings hook escape only when that hook is absent — i.e. by being ungated.

This adds the client. It implements protocol steps 1, 4 and 5:

  1. begin_verification  -> mint a challenge, return the exact receipt text
  2. the MODEL emits that text verbatim as its entire latest message
  3. the runtime adapter ships that message to the observer socket
  4. observe_receipt      -> PENDING_PROMOTION
  5. promote_lease        -> VERIFIED

Step 2 is deliberately NOT implemented here, and that is the point.
is_verbatim_receipt uses hmac.compare_digest against the exact minted string —
explicitly "not a transcript substring" — which makes promotion a LIVENESS
PROOF: it requires a live model that received the challenge in its context and
echoed it exactly.

receipt-observer-client.py will post ANY string as the latest assistant message.
A promotion client that posted its own receipt would satisfy the broker while
proving nothing — a gate-disabler indistinguishable from a working fix unless
someone specifically looks. Emitting the receipt therefore belongs to the runtime
adapter, where a real model turn happens. A local diagnostic that posts its own
receipt exists in the operator's repo and is deliberately NOT shipped here.

The construction binds the exact normative source bytes, so a VERIFIED lease
means "this agent is running THIS law", not merely "this session id is known".
h_source/h_payload are derived by importing the framework's own
normative_fragments.build_payload rather than reimplementing it: the broker
derives them the same way and any divergence yields PAYLOAD_BINDING_MISMATCH.
There must be exactly one implementation.

session_identity() prefers the generation FILE over the env var, matching
lease_generation.py. Sending a generation higher than the broker's would revoke
the session's own authority (daemon.py:342-344), so it never guesses.

Verified end-to-end on sb-it-1-dt under a real lease-gated anchor: a mutator
denied rc=2 MUTATOR_UNVERIFIED, then begin -> observe -> promote -> VERIFIED,
then the same mutator allowed rc=0. Negative controls pass: a fresh session is
still denied, and an unrelated session still reads UNVERIFIED — promotion is
per-session and does not leak.

Still open: adapter wiring for step 2. Lazy promotion on first mutator attempt
avoids colliding with the Constitution's first-response mode declaration, since
compare_digest requires the receipt to be the WHOLE message.
2026-08-11 17:03:58 -05:00
Jason Woltjeandmos-dt-0 65cafe3e9e feat(launch): isolate harness homes and record immutable launch provenance
Mosaic wrote into the operator's harness base installs — ~/.claude,
~/.pi/agent, ~/.codex, ~/.config/opencode — for settings, instructions, and a
102-symlink skill farm per harness. Any experiment with hooks or gating
therefore mutated the operator's own tooling, and a broken framework change
could take out the very harness needed to repair it.

Harness home isolation
----------------------
Each runtime now reads config from a dedicated mosaic-owned home via the
harness's own config-dir variable:

  claude    CLAUDE_CONFIG_DIR     ~/.config/mosaic/.claude
  pi        PI_CODING_AGENT_DIR   ~/.config/mosaic/.pi     (replaces ~/.pi/agent)
  codex     CODEX_HOME            ~/.config/mosaic/.codex
  opencode  XDG_CONFIG_HOME       ~/.config/mosaic/.opencode

These paths are manifest-UNKNOWN, so rule 3 (#791) resolves them to operator
ownership and a keep-mode upgrade can neither overwrite nor prune them.
A bare `claude` / `pi` keeps its own config AND auth, making it a structural
break-glass rather than one depending on restoring a file under pressure.

opencode is blunter than the rest: it has no dedicated variable and follows XDG,
so isolation also relocates XDG lookups for anything it spawns. Documented in
place.

mosaic-sync-skills now links into those homes and cleans the legacy farms it
previously planted in base installs. Ownership is proven by RESOLUTION, not by
name — only symlinks resolving inside the canonical/local skills dirs are
removed, mirroring the refusal already in commands/skill.js. Verified against a
real install: codex's own .system directory survived while its 102 mosaic links
were removed. Both resolution prefixes are length-checked first; an empty prefix
would make "$resolved" == "$prefix/"* match every absolute path and delete
foreign symlinks.

Immutable launch record
-----------------------
Every launch now appends one record to fleet/run/sessions/events.ndjson before
exec. Mandatory, mechanical, no model involvement.

pi rewrites its own argv to a bare `pi`, so /proc/<pid>/cmdline destroys the
launch evidence — that has already produced a confident wrong diagnosis ("this
agent bypassed the launcher"), disproved only by the parent's argv and only
because the parent had not yet exited. A record written before exec is the only
place this survives.

The path is the #797 Runtime Session Ledger, already operator-classified and
already covered by test-upgrade-manifest-guard.sh, which seeds it and proves a
populated ledger survives keep-mode upgrades — but nothing shipped ever wrote
it. This implements it in the shape that guard already asserts (0600 files under
a 0700 dir).

`mosaic` writes session.launch; launch-runtime.py appends lease.register with
the broker session id and activation capability. They correlate by an explicit
MOSAIC_LAUNCH_ID, never by pid: execRuntime uses spawnSync, so the runtime is a
child with a different pid.

Records normative fragment digests (CONSTITUTION/AGENTS/SOUL/USER/STANDARDS/
TOOLS/RUNTIME) — the same set the broker hashes for promotion, so drift is
mechanically detectable rather than a matter of judgement.

Credential-safe: env is captured as PRESENT NAMES ONLY, and argv values over
256 bytes become a sha256 + length rather than being inlined.

Also fixes CLI_VERSION resolution: '@mosaicstack/mosaic/package.json' is not in
the package exports map and always throws ERR_PACKAGE_PATH_NOT_EXPORTED.
resolveTool() uses that same failing specifier, which is why its documented
preference for bundled tools over the deployed ~/.config/mosaic copy has never
once applied — noted in place, not fixed here.

Verified on sb-it-1-dt: isolated homes written and base installs byte-identical
for all four harnesses; 408 legacy symlinks removed with 1 foreign entry
preserved; launch records paired across the spawn boundary. typecheck shows zero
errors in launch.ts (the @mosaicstack/types failures are pre-existing and
reproduce on a pristine origin/main worktree).
2026-08-11 17:03:58 -05:00
be-coder-08andmos-dt-0 5ea64a85cf feat(pr-merge): preserve linked authors in squash messages (#1066)
Co-authored-by: be-coder-08 <[email protected]>
2026-08-11 17:03:57 -05:00
be-coder-08andmos-dt-0 f6e81f8dc9 fix(ci): remove upgrade rollback signal race (#1060)
Co-authored-by: be-coder-08 <[email protected]>
2026-08-11 17:03:26 -05:00
be-coder-08andmos-dt-0 d0dbbab753 test(ci): make queue guard harness deterministic (#1062)
Co-authored-by: be-coder-08 <[email protected]>
2026-08-11 17:03:26 -05:00
mos-dt-0 0e4eb1445c Merge pull request 'fix(installer): propagate wizard gateway failures' (#1134) from fix/wizard-gateway-failure into next
ci/woodpecker/push/publish Pipeline failed
2026-08-11 22:01:43 +00:00
mos-dt-0 592d60425f Merge pull request 'fix(installer): require Node 22 for the --next lane' (#1130) from fix/next-node-gate into next
ci/woodpecker/push/publish Pipeline was canceled
2026-08-11 22:01:39 +00:00
mos-dt-0 a43f343efd Merge pull request 'fix(security): mosaic-init RCE — eval on prompt answers → printf -v (#1115)' (#1127) from fix/mosaic-init-rce into next
ci/woodpecker/push/publish Pipeline was canceled
2026-08-11 22:01:35 +00:00
jason.woltje 88ef9d4fa5 Merge pull request 'P3-R1: repair routing health-enum (#1) + wire /mcp command (#5)' (#1154) from feat/webui-p3r1-routing-mcp into next
ci/woodpecker/push/publish Pipeline failed
2026-08-11 20:51:52 +00:00
shaggy (mosaic-dev box) bda308efd9 fix(gateway): repair routing health and MCP command wiring
ci/woodpecker/pr/ci Pipeline was successful
2026-08-11 15:28:57 -05:00
jason.woltje 20718b5a27 Merge pull request 'P4-1: read-only Projects + Tasks SPA pages + dev-port pins' (#1153) from feat/webui-p4-1 into next
ci/woodpecker/push/publish Pipeline failed
2026-08-11 04:00:36 +00:00
shaggy (mosaic-dev box) 29db24210c fix(web): P4-1 restore graceful degradation for missions/tasks fetch on project detail
ci/woodpecker/pr/ci Pipeline was successful
2026-08-10 22:43:47 -05:00
shaggy (mosaic-dev box)andClaude Haiku 4.5 a6085eea37 feat(web): add read-only project and task SPA pages
Co-Authored-By: Claude Haiku 4.5 <[email protected]>
2026-08-10 22:23:59 -05:00
Mos e00cc475a2 Merge pull request 'P3 — Typed SPA chat (Phase P webUI)' (#1151) from feat/webui-p3-chat into next
ci/woodpecker/push/publish Pipeline failed
2026-08-10 22:57:07 +00:00
shaggy (mosaic-dev box) 7d84e4ee03 fix(gateway): sanitize raw exceptions in /mcp status + /reload sources (P3 re-review#5 blocker)
ci/woodpecker/pr/ci Pipeline was successful
2026-08-10 16:58:07 -05:00
shaggy (mosaic-dev box) 4aaf41dd1a fix(web,gateway): close P3 re-review#4 findings — sanitize all executor catches; lock pre-start turn boundary (wire turnId deferred) 2026-08-10 16:07:12 -05:00
mos-dt-0 bf32f29acd fix(ci): make queue guard purpose-sensitive (#1148)
ci/woodpecker/push/publish Pipeline failed
2026-08-10 20:54:09 +00:00
Jason Woltje 1655b1579a Move deploy and briefs into docs folder. Remove old files.
ci/woodpecker/push/publish Pipeline was canceled
2026-08-10 14:44:19 -05:00
Jason Woltje e478a359eb Move briefs directory inside the docs
ci/woodpecker/push/publish Pipeline was canceled
2026-08-10 14:42:04 -05:00
Jason Woltje 76e4242cb1 Reformat Requirements block
ci/woodpecker/push/publish Pipeline was canceled
2026-08-10 14:40:18 -05:00
Jason Woltje 00eb216480 docs(agents): consolidate project guidance 2026-08-10 14:27:59 -05:00
shaggy (mosaic-dev box) d46a2d675a fix(web,gateway): close P3 re-review#3 findings — sanitize command errors, harden turn-lock & caps 2026-08-10 14:21:52 -05:00
shaggy (mosaic-dev box) 48bb19310d fix(web): close P3 chat re-review findings 2026-08-10 01:58:27 -05:00
shaggy (mosaic-dev box)andClaude Haiku 4.5 caebf9ef70 fix(web): harden typed SPA chat lifecycle
Co-Authored-By: Claude Haiku 4.5 <[email protected]>
2026-08-10 00:24:20 -05:00
shaggy (mosaic-dev box) b2e005f2b4 feat(web): add typed SPA chat
Bring the chat experience into the Vite/React-Router SPA on the exact typed
Socket.IO /chat contract from @mosaicstack/types, replacing the /chat
placeholder behind AuthGuard. Surfaces message:ack (with an accessible
status), agent:start, streamed agent:text/agent:thinking, tool start/end
status, agent:end with usage, session:info (thinking controls + routing
decision), commands:manifest, command:result, command:approval (with a
one-time approved-run affordance), system:reload (refreshing the rendered
manifest), and error, and emits message/abort/set:thinking/command:execute/
command:approve with exact payloads.

The gateway does not guarantee message:ack is the first event for a new
conversation (session:info, and error on auth/session-creation failure, can
both arrive first) — conversation-scoped events now adopt the conversation
from whichever scoped event names it first while a send is pending, then
filter everything else against that established conversation. A typed error
stops streaming instead of leaving Stop stuck active; agent:end no longer
appends an empty assistant turn when there is no text or thinking; and a
second message can no longer be sent while a turn is streaming.

Command approval is now integrity-checked end to end: only one
command:approve request may be outstanding at a time (a concurrent request
is ignored rather than overwriting the pending command/args), a stale or
mismatched command:approval response cannot replace active approval state,
and running an approved command clears its approval state immediately (via
a ref, before React re-renders) so a double-click cannot replay
command:execute.

The `/chat` socket is now typed at a single boundary: apps/web/src/lib/
socket.ts narrows socket.io-client's untyped `io()` return value to
`ChatSocket` (Socket<ServerToClientEvents, ClientToServerEvents>) once, at
creation, via the one assertion the library's types force; every consumer
(use-chat-connection.ts) then gets fully checked `on`/`emit` calls with no
further casts. The shared contract types live in the new
apps/web/src/lib/chat-contract.ts (replacing the old spa/chat/types.ts
shim), which re-exports them via type-only imports resolved directly
against packages/types/src (apps/web has no @mosaicstack/types package
dependency, so this stays source-only and is erased at compile time —
no package manifest or lockfile is touched). The two recorded-event test
suites now drive a shared, typed fake socket
(spa/chat/test-support/fake-chat-socket.ts) instead of an untyped
`(event: string, payload: unknown)` harness, so a wrong event name or
malformed payload fails to compile.
2026-08-09 21:39:16 -05:00
Jason Woltje dc67590a96 fix(installer): propagate wizard gateway failures (#1120)
ci/woodpecker/pr/ci Pipeline was successful
2026-08-09 05:01:09 -05:00
Jason Woltje baf4306f51 fix(installer): require Node 22 for next lane
ci/woodpecker/pr/ci Pipeline was successful
2026-08-09 02:37:03 -05:00
Jason Woltje 12677a928d fix(mosaic): prevent init prompt code execution
ci/woodpecker/pr/ci Pipeline was successful
2026-08-09 00:28:57 -05:00
100 changed files with 12671 additions and 377 deletions
+62 -23
View File
@@ -11,48 +11,87 @@
## Project Context ## Project Context
Mosaic Stack is a self-hosted, multi-user AI agent platform. TypeScript monorepo with NestJS gateway, Next.js web dashboard, Pi SDK agent runtime, and plugin architecture for Discord/Telegram. Mosaic Stack is a self-hosted, multi-user AI agent platform. It is a TypeScript monorepo with a NestJS gateway, Next.js dashboard, Pi SDK agent runtime, and Discord/Telegram plugin architecture.
## Package Map ### Stack
- **API:** NestJS with Fastify (`apps/gateway`)
- **Web:** Next.js 16 with React 19 (`apps/web`)
- **ORM and database:** Drizzle ORM, PostgreSQL 17, and pgvector (`packages/db`)
- **Authentication:** BetterAuth (`packages/auth`)
- **Agent runtime:** Pi SDK (`apps/gateway`, `packages/mosaic`)
- **Queue:** Valkey 8 (`packages/queue`)
- **Build:** pnpm workspaces and Turborepo
- **CI:** Woodpecker CI
- **Observability:** OpenTelemetry and Jaeger
### Package Map
| Package | Purpose | Key Dependencies | | Package | Purpose | Key Dependencies |
| ------------------ | ------------------------------- | -------------------------------- | | ------------------ | ----------------------------- | -------------------------------- |
| `apps/gateway` | NestJS API + WebSocket hub | Fastify, Socket.IO, Pi SDK, OTEL | | `apps/gateway` | NestJS API + WebSocket hub | Fastify, Socket.IO, Pi SDK, OTEL |
| `apps/web` | Next.js dashboard | React 19, Tailwind | | `apps/web` | Next.js dashboard | React 19, Tailwind |
| `packages/types` | Shared TypeScript contracts | class-validator | | `packages/types` | Shared TypeScript contracts | class-validator |
| `packages/db` | Drizzle ORM schema + migrations | drizzle-orm, postgres | | `packages/db` | Drizzle schema and migrations | drizzle-orm, postgres |
| `packages/auth` | BetterAuth configuration | better-auth, @mosaicstack/db | | `packages/auth` | BetterAuth configuration | better-auth, @mosaicstack/db |
| `packages/brain` | Data layer (PG-backed) | @mosaicstack/db | | `packages/brain` | Structured data layer | @mosaicstack/db |
| `packages/queue` | Valkey task queue + MCP | ioredis | | `packages/queue` | Valkey task queue and MCP | ioredis |
| `packages/coord` | Mission coordination | @mosaicstack/queue | | `packages/coord` | Mission coordination | @mosaicstack/queue |
| `packages/mosaic` | Unified `mosaic` CLI + TUI | Ink, Pi SDK, commander | | `packages/mosaic` | Unified `mosaic` CLI and TUI | Ink, Pi SDK, commander |
| `plugins/discord` | Discord channel plugin | discord.js | | `plugins/discord` | Discord channel plugin | discord.js |
| `plugins/telegram` | Telegram channel plugin | Telegraf | | `plugins/telegram` | Telegram channel plugin | Telegraf |
## Architecture Rules ## Architecture and Code Conventions
1. Gateway is the single API surface all clients connect through it 1. Gateway is the single API surface; all clients connect through it.
2. Pi SDK is ESM-only gateway and CLI must use ESM 2. Pi SDK is ESM-only; gateway and CLI code must remain ESM.
3. Socket.IO typed events defined in `@mosaicstack/types` enforce compile-time contracts 3. Use `"type": "module"`, NodeNext module resolution, and `.js` extensions in imports.
4. OTEL auto-instrumentation loads before NestJS bootstrap 4. Keep typed Socket.IO events in `@mosaicstack/types` to enforce client/server contracts.
5. BetterAuth manages auth tables; schema defined in `@mosaicstack/db` 5. Import OTEL tracing before NestJS bootstrap (`import './tracing.js'`).
6. Docker Compose provides PG (5433), Valkey (6380), OTEL Collector (4317/4318), Jaeger (16686) 6. Use explicit `@Inject()` decorators in NestJS because tsx/esbuild does not emit decorator metadata.
7. Explicit `@Inject()` decorators required in NestJS (tsx/esbuild doesn't emit decorator metadata) 7. Keep DTOs in `*.dto.ts` files at module boundaries.
8. BetterAuth owns authentication tables; their schema is defined in `@mosaicstack/db`.
9. Create a task-specific scratchpad for non-trivial work.
## Development Workflow ## Development Workflow
Requirements: Node.js 20+, pnpm 10.6.2, and Docker Compose when optional local services are needed.
```bash ```bash
docker compose up -d # Infrastructure pnpm install --frozen-lockfile
pnpm install # Dependencies pnpm preflight
pnpm typecheck && pnpm lint && pnpm format:check # Quality gates
# Optional local queue service only; do not start the full Compose stack.
docker compose up -d valkey
``` ```
## Repo-Specific Notes The pre-push hook requires:
- DTOs in `*.dto.ts` files at module boundaries ```bash
- ESM everywhere (`"type": "module"`, `.js` extensions in imports) pnpm preflight && pnpm typecheck && pnpm lint && pnpm format:check
- NodeNext module resolution in all tsconfigs ```
- Scratchpads are mandatory for non-trivial tasks
Software delivery also requires the applicable tests. Common repository commands are:
```bash
pnpm typecheck # TypeScript checks across the workspace
pnpm lint # ESLint across the workspace
pnpm test # Checkout tests and package Vitest suites
pnpm format:check # Prettier check
pnpm build # Build all packages and applications
```
## Database and Local Runtime Safety
- Current local data-layer work uses in-process PGlite; leave `DATABASE_URL` unset.
- PostgreSQL execution is held until KBN-101-00, KBN-101-03, and KBN-101-05 land.
- Do not invoke a migration runner, initialization SQL, or the Compose PostgreSQL service from this checkout.
- Do not start Gateway/Web or run root `pnpm dev` as a local PGlite route. The current dotenv loader can inherit a daemon PostgreSQL DSN; KBN-101-02 must make that path fail closed first.
- Migration artifact generation is offline and does not authorize PostgreSQL access:
```bash
pnpm --filter @mosaicstack/db db:generate
```
## docs/TASKS.md — Schema (CANONICAL) ## docs/TASKS.md — Schema (CANONICAL)
+3 -44
View File
@@ -1,46 +1,5 @@
# CLAUDE.md — Mosaic Stack # Claude Compatibility Pointer
## Project @AGENTS.md
Self-hosted, multi-user AI agent platform. TypeScript monorepo. Do not add project guidance here. Keep `AGENTS.md` authoritative so every agent runtime receives the same instructions.
## Stack
- **API**: NestJS + Fastify adapter (`apps/gateway`)
- **Web**: Next.js 16 + React 19 (`apps/web`)
- **ORM**: Drizzle ORM + PostgreSQL 17 + pgvector (`packages/db`)
- **Auth**: BetterAuth (`packages/auth`)
- **Agent**: Pi SDK (`packages/agent`, `packages/mosaic`)
- **Queue**: Valkey 8 (`packages/queue`)
- **Build**: pnpm workspaces + Turborepo
- **CI**: Woodpecker CI
- **Observability**: OpenTelemetry → Jaeger
## Commands
```bash
pnpm typecheck # TypeScript check (all packages)
pnpm lint # ESLint (all packages)
pnpm format:check # Prettier check
pnpm test # Vitest (all packages)
pnpm build # Build all packages
# Database
pnpm --filter @mosaicstack/db db:generate # Offline migration artifact generation only
# PostgreSQL execution is held until KBN-101-00/-03/-05 land. Do not invoke a runner,
# init SQL, or Compose PostgreSQL service from this checkout.
# Dev: local PGlite data-layer work needs no PostgreSQL. Optional local queue service only:
docker compose up -d valkey
# Do not start Gateway/Web or root pnpm dev as a local PGlite route: the current unguarded dotenv
# loader can inherit a daemon PostgreSQL DSN. KBN-101-02 must make that state fail closed first.
```
## Conventions
- ESM everywhere (`"type": "module"`, `.js` extensions in imports)
- NodeNext module resolution
- Explicit `@Inject()` decorators in NestJS (tsx/esbuild doesn't support emitDecoratorMetadata)
- DTOs in `*.dto.ts` files at module boundaries
- OTEL tracing imported before NestJS bootstrap (`import './tracing.js'`)
- All three gates must pass before push: typecheck, lint, format:check
+7 -3
View File
@@ -48,9 +48,13 @@ mosaic wizard # Full guided setup (gateway install → verify)
### Requirements ### Requirements
- Node.js ≥ 20 - Node.js ≥ 22
- npm (for global @mosaicstack/mosaic install) - npm (for global @mosaicstack/mosaic install)
- One or more runtimes: [Claude Code](https://docs.anthropic.com/en/docs/claude-code), [Codex](https://github.com/openai/codex), [OpenCode](https://opencode.ai), or [Pi](https://github.com/mariozechner/pi-coding-agent) - One or more runtimes:
- [Claude Code](https://docs.anthropic.com/en/docs/claude-code)
- [Codex](https://github.com/openai/codex)
- [OpenCode](https://opencode.ai)
- [Pi](https://pi.dev)
## Usage ## Usage
@@ -200,7 +204,7 @@ Consent state is persisted in config. Remote upload is a no-op until you run `mo
### Prerequisites ### Prerequisites
- Node.js ≥ 20 - Node.js ≥ 22
- pnpm 10.6+ - pnpm 10.6+
- Docker & Docker Compose - Docker & Docker Compose
@@ -8,6 +8,7 @@
* to avoid real I/O — they verify the complete classify → match → decide path. * to avoid real I/O — they verify the complete classify → match → decide path.
*/ */
import { describe, it, expect, vi } from 'vitest'; import { describe, it, expect, vi } from 'vitest';
import type { ProviderHealthStatus } from '@mosaicstack/types';
import { RoutingEngineService } from './routing-engine.service.js'; import { RoutingEngineService } from './routing-engine.service.js';
import { DEFAULT_ROUTING_RULES } from '../routing/default-rules.js'; import { DEFAULT_ROUTING_RULES } from '../routing/default-rules.js';
import type { RoutingRule } from './routing.types.js'; import type { RoutingRule } from './routing.types.js';
@@ -17,7 +18,7 @@ import type { RoutingRule } from './routing.types.js';
/** Build a RoutingEngineService backed by the given rule set and health map. */ /** Build a RoutingEngineService backed by the given rule set and health map. */
function makeService( function makeService(
rules: RoutingRule[], rules: RoutingRule[],
healthMap: Record<string, { status: string }>, healthMap: Record<string, { status: ProviderHealthStatus }>,
): RoutingEngineService { ): RoutingEngineService {
const mockDb = { const mockDb = {
select: vi.fn().mockReturnValue({ select: vi.fn().mockReturnValue({
@@ -67,11 +68,11 @@ function defaultRules(): RoutingRule[] {
} }
/** A health map where anthropic, openai, and zai are all healthy. */ /** A health map where anthropic, openai, and zai are all healthy. */
const allHealthy: Record<string, { status: string }> = { const allHealthy: Record<string, { status: ProviderHealthStatus }> = {
anthropic: { status: 'up' }, anthropic: { status: 'healthy' },
openai: { status: 'up' }, openai: { status: 'healthy' },
zai: { status: 'up' }, zai: { status: 'healthy' },
ollama: { status: 'up' }, ollama: { status: 'healthy' },
}; };
// ─── M4-013 E2E tests ───────────────────────────────────────────────────────── // ─── M4-013 E2E tests ─────────────────────────────────────────────────────────
@@ -212,10 +213,10 @@ describe('M4-013: routing end-to-end pipeline', () => {
// Let's use a simple coding message to target Simple coding → Codex (openai) // Let's use a simple coding message to target Simple coding → Codex (openai)
const message = 'implement a sort function'; const message = 'implement a sort function';
const unhealthyHealth = { const unhealthyHealth: Record<string, { status: ProviderHealthStatus }> = {
anthropic: { status: 'down' }, anthropic: { status: 'down' },
openai: { status: 'up' }, openai: { status: 'healthy' },
zai: { status: 'up' }, zai: { status: 'healthy' },
ollama: { status: 'down' }, ollama: { status: 'down' },
}; };
@@ -1,5 +1,6 @@
import { Inject, Injectable, Logger } from '@nestjs/common'; import { Inject, Injectable, Logger } from '@nestjs/common';
import { routingRules, type Db, and, asc, eq, or } from '@mosaicstack/db'; import { routingRules, type Db, and, asc, eq, or } from '@mosaicstack/db';
import type { ProviderHealthStatus } from '@mosaicstack/types';
import { DB } from '../../database/database.module.js'; import { DB } from '../../database/database.module.js';
import { ProviderService } from '../provider.service.js'; import { ProviderService } from '../provider.service.js';
import { classifyTask } from './task-classifier.js'; import { classifyTask } from './task-classifier.js';
@@ -49,7 +50,7 @@ export class RoutingEngineService {
async resolve( async resolve(
message: string, message: string,
userId?: string, userId?: string,
availableProviders?: Record<string, { status: string }>, availableProviders?: Record<string, { status: ProviderHealthStatus }>,
): Promise<RoutingDecision> { ): Promise<RoutingDecision> {
const classification = classifyTask(message); const classification = classifyTask(message);
this.logger.debug( this.logger.debug(
@@ -69,9 +70,8 @@ export class RoutingEngineService {
if (!this.matchConditions(rule, classification)) continue; if (!this.matchConditions(rule, classification)) continue;
const providerStatus = health[rule.action.provider]?.status; const providerStatus = health[rule.action.provider]?.status;
const isHealthy = providerStatus === 'up' || providerStatus === 'ok';
if (!isHealthy) { if (!this.isRoutable(providerStatus)) {
this.logger.debug( this.logger.debug(
`Rule "${rule.name}" matched but provider "${rule.action.provider}" is unhealthy (status: ${providerStatus ?? 'unknown'})`, `Rule "${rule.name}" matched but provider "${rule.action.provider}" is unhealthy (status: ${providerStatus ?? 'unknown'})`,
); );
@@ -111,6 +111,10 @@ export class RoutingEngineService {
// ─── Private helpers ─────────────────────────────────────────────────────── // ─── Private helpers ───────────────────────────────────────────────────────
private isRoutable(status: ProviderHealthStatus | undefined): boolean {
return status === 'healthy' || status === 'degraded';
}
private evaluateCondition( private evaluateCondition(
condition: RoutingCondition, condition: RoutingCondition,
classification: TaskClassification, classification: TaskClassification,
@@ -186,11 +190,12 @@ export class RoutingEngineService {
* Walk the fallback chain and return the first healthy provider/model pair. * Walk the fallback chain and return the first healthy provider/model pair.
* If none are healthy, return the first entry unconditionally (last resort). * If none are healthy, return the first entry unconditionally (last resort).
*/ */
private applyFallbackChain(health: Record<string, { status: string }>): RoutingDecision { private applyFallbackChain(
health: Record<string, { status: ProviderHealthStatus }>,
): RoutingDecision {
for (const candidate of FALLBACK_CHAIN) { for (const candidate of FALLBACK_CHAIN) {
const providerStatus = health[candidate.provider]?.status; const providerStatus = health[candidate.provider]?.status;
const isHealthy = providerStatus === 'up' || providerStatus === 'ok'; if (this.isRoutable(providerStatus)) {
if (isHealthy) {
this.logger.debug(`Fallback resolved: ${candidate.provider}/${candidate.model}`); this.logger.debug(`Fallback resolved: ${candidate.provider}/${candidate.model}`);
return { return {
provider: candidate.provider, provider: candidate.provider,
@@ -1,4 +1,5 @@
import { describe, it, expect, vi, beforeEach } from 'vitest'; import { describe, it, expect, vi, beforeEach } from 'vitest';
import type { ProviderHealthStatus } from '@mosaicstack/types';
import { RoutingEngineService } from './routing-engine.service.js'; import { RoutingEngineService } from './routing-engine.service.js';
import type { RoutingRule, TaskClassification } from './routing.types.js'; import type { RoutingRule, TaskClassification } from './routing.types.js';
@@ -29,7 +30,7 @@ function makeClassification(overrides: Partial<TaskClassification> = {}): TaskCl
/** Build a minimal RoutingEngineService with mocked DB and ProviderService. */ /** Build a minimal RoutingEngineService with mocked DB and ProviderService. */
function makeService( function makeService(
rules: RoutingRule[] = [], rules: RoutingRule[] = [],
healthMap: Record<string, { status: string }> = {}, healthMap: Record<string, { status: ProviderHealthStatus }> = {},
): RoutingEngineService { ): RoutingEngineService {
const mockDb = { const mockDb = {
select: vi.fn().mockReturnValue({ select: vi.fn().mockReturnValue({
@@ -217,7 +218,10 @@ describe('RoutingEngineService.resolve — priority ordering', () => {
}), }),
]; ];
const service = makeService(rules, { anthropic: { status: 'up' }, openai: { status: 'up' } }); const service = makeService(rules, {
anthropic: { status: 'healthy' },
openai: { status: 'healthy' },
});
const decision = await service.resolve('implement a function'); const decision = await service.resolve('implement a function');
expect(decision.ruleName).toBe('high priority'); expect(decision.ruleName).toBe('high priority');
@@ -241,7 +245,10 @@ describe('RoutingEngineService.resolve — priority ordering', () => {
}), }),
]; ];
const service = makeService(rules, { anthropic: { status: 'up' }, openai: { status: 'up' } }); const service = makeService(rules, {
anthropic: { status: 'healthy' },
openai: { status: 'healthy' },
});
const decision = await service.resolve('implement a function'); const decision = await service.resolve('implement a function');
expect(decision.ruleName).toBe('coding rule'); expect(decision.ruleName).toBe('coding rule');
@@ -270,7 +277,7 @@ describe('RoutingEngineService.resolve — unhealthy provider handling', () => {
const service = makeService(rules, { const service = makeService(rules, {
anthropic: { status: 'down' }, // primary is unhealthy anthropic: { status: 'down' }, // primary is unhealthy
openai: { status: 'up' }, openai: { status: 'healthy' },
}); });
const decision = await service.resolve('implement a function'); const decision = await service.resolve('implement a function');
@@ -290,7 +297,7 @@ describe('RoutingEngineService.resolve — unhealthy provider handling', () => {
]; ];
const service2 = makeService(unhealthyRules, { const service2 = makeService(unhealthyRules, {
anthropic: { status: 'up' }, anthropic: { status: 'healthy' },
openai: { status: 'down' }, openai: { status: 'down' },
}); });
@@ -306,7 +313,7 @@ describe('RoutingEngineService.resolve — unhealthy provider handling', () => {
const service = makeService(rules, { const service = makeService(rules, {
anthropic: { status: 'down' }, // Sonnet is on anthropic — down anthropic: { status: 'down' }, // Sonnet is on anthropic — down
ollama: { status: 'up' }, // Haiku is also on anthropic — use Ollama as next ollama: { status: 'healthy' }, // Haiku is also on anthropic — use Ollama as next
}); });
const decision = await service.resolve('hello there'); const decision = await service.resolve('hello there');
@@ -345,7 +352,7 @@ describe('RoutingEngineService.resolve — empty conditions (fallback rule)', ()
}), }),
]; ];
const service = makeService(rules, { anthropic: { status: 'up' } }); const service = makeService(rules, { anthropic: { status: 'healthy' } });
const decision = await service.resolve('completely unrelated message xyz'); const decision = await service.resolve('completely unrelated message xyz');
expect(decision.ruleName).toBe('catch-all'); expect(decision.ruleName).toBe('catch-all');
@@ -369,7 +376,7 @@ describe('RoutingEngineService.resolve — empty conditions (fallback rule)', ()
}), }),
]; ];
const service = makeService(rules, { anthropic: { status: 'up' } }); const service = makeService(rules, { anthropic: { status: 'healthy' } });
const codingDecision = await service.resolve('implement a function'); const codingDecision = await service.resolve('implement a function');
expect(codingDecision.ruleName).toBe('specific coding rule'); expect(codingDecision.ruleName).toBe('specific coding rule');
@@ -401,7 +408,7 @@ describe('RoutingEngineService.resolve — disabled rules', () => {
}), }),
]; ];
const service = makeService(rules, { anthropic: { status: 'up' } }); const service = makeService(rules, { anthropic: { status: 'healthy' } });
const decision = await service.resolve('implement a function'); const decision = await service.resolve('implement a function');
expect(decision.ruleName).toBe('enabled fallback'); expect(decision.ruleName).toBe('enabled fallback');
@@ -452,9 +459,45 @@ describe('RoutingEngineService.resolve — availableProviders override', () => {
ps: unknown, ps: unknown,
) => RoutingEngineService)(mockDb, mockProviderService); ) => RoutingEngineService)(mockDb, mockProviderService);
const preSupplied = { anthropic: { status: 'up' } }; const preSupplied: Record<string, { status: ProviderHealthStatus }> = {
anthropic: { status: 'healthy' },
};
await service.resolve('implement a function', undefined, preSupplied); await service.resolve('implement a function', undefined, preSupplied);
expect(mockHealthCheckAll).not.toHaveBeenCalled(); expect(mockHealthCheckAll).not.toHaveBeenCalled();
}); });
}); });
// ─── resolve — canonical ProviderHealthStatus values ──────────────────────────
describe('RoutingEngineService.resolve — canonical health status routing', () => {
it('routes healthy and degraded providers by rule, and falls through to fallback when down', async () => {
const codingRule = makeRule({
name: 'coding rule',
priority: 1,
conditions: [{ field: 'taskType', operator: 'eq', value: 'coding' }],
action: { provider: 'openai', model: 'gpt-4o' },
});
// healthy → selected by its own rule, not the fallback chain
const healthyService = makeService([codingRule], { openai: { status: 'healthy' } });
const healthyDecision = await healthyService.resolve('implement a function');
expect(healthyDecision.ruleName).toBe('coding rule');
expect(healthyDecision.provider).toBe('openai');
// down → rule is skipped as unroutable, falls through to the fallback chain
const downService = makeService([codingRule], {
openai: { status: 'down' },
anthropic: { status: 'healthy' },
});
const downDecision = await downService.resolve('implement a function');
expect(downDecision.ruleName).toBe('fallback');
expect(downDecision.provider).toBe('anthropic');
// degraded → still routable, selected by its own rule, not the fallback chain
const degradedService = makeService([codingRule], { openai: { status: 'degraded' } });
const degradedDecision = await degradedService.resolve('implement a function');
expect(degradedDecision.ruleName).toBe('coding rule');
expect(degradedDecision.provider).toBe('openai');
});
});
@@ -1,3 +1,4 @@
import { Logger } from '@nestjs/common';
import { describe, it, expect, vi, beforeEach } from 'vitest'; import { describe, it, expect, vi, beforeEach } from 'vitest';
import { CommandExecutorService } from './command-executor.service.js'; import { CommandExecutorService } from './command-executor.service.js';
import type { SlashCommandPayload } from '@mosaicstack/types'; import type { SlashCommandPayload } from '@mosaicstack/types';
@@ -12,6 +13,7 @@ const mockRegistry = {
{ name: 'agent', aliases: ['a'], scope: 'agent', execution: 'socket', available: true }, { name: 'agent', aliases: ['a'], scope: 'agent', execution: 'socket', available: true },
{ name: 'prdy', aliases: [], scope: 'agent', execution: 'socket', available: true }, { name: 'prdy', aliases: [], scope: 'agent', execution: 'socket', available: true },
{ name: 'tools', aliases: [], scope: 'agent', execution: 'socket', available: true }, { name: 'tools', aliases: [], scope: 'agent', execution: 'socket', available: true },
{ name: 'mcp', aliases: [], scope: 'agent', execution: 'socket', available: true },
], ],
skills: [], skills: [],
})), })),
@@ -72,7 +74,20 @@ const mockChatGateway = {
broadcastSessionInfo: vi.fn(), broadcastSessionInfo: vi.fn(),
}; };
function buildService(redis: typeof mockRedis | null = mockRedis): CommandExecutorService { const mockMcpClient = {
reconnectServer: vi.fn().mockResolvedValue(undefined),
getServerStatuses: vi.fn(() => []),
getToolDefinitions: vi.fn(() => []),
};
function buildService(
redis: typeof mockRedis | null = mockRedis,
mcpClient: {
reconnectServer: ReturnType<typeof vi.fn>;
getServerStatuses: ReturnType<typeof vi.fn>;
getToolDefinitions: ReturnType<typeof vi.fn>;
} = mockMcpClient,
): CommandExecutorService {
return new CommandExecutorService( return new CommandExecutorService(
mockRegistry as never, mockRegistry as never,
mockAgentService as never, mockAgentService as never,
@@ -82,7 +97,7 @@ function buildService(redis: typeof mockRedis | null = mockRedis): CommandExecut
mockBrain as never, mockBrain as never,
null, null,
mockChatGateway as never, mockChatGateway as never,
null, mcpClient as never,
); );
} }
@@ -258,4 +273,124 @@ describe('CommandExecutorService — P8-012 commands', () => {
expect(result.command).toBe('tools'); expect(result.command).toBe('tools');
expect(result.message).toContain('tools'); expect(result.message).toContain('tools');
}); });
// Top-level catch sanitization (P3-4 re-review finding #1): a rejected
// Redis `set` inside /provider login is the only reachable path into the
// top-level catch in `execute()`. The raw exception must be logged
// server-side but never handed back to the socket client.
it('sanitizes the top-level command catch, logging the raw exception but never returning it to the client', async () => {
const distinctiveRawFailure = 'ECONNREFUSED distinctive-raw-redis-failure-token-9f31';
const rawError = new Error(distinctiveRawFailure);
const failingRedis = {
set: vi.fn().mockRejectedValue(rawError),
get: vi.fn(),
del: vi.fn(),
};
const failingService = buildService(failingRedis as unknown as typeof mockRedis);
const loggerErrorSpy = vi.spyOn(Logger.prototype, 'error').mockImplementation(() => undefined);
const payload: SlashCommandPayload = {
command: 'provider',
args: 'login anthropic',
conversationId,
};
const result = await failingService.execute(payload, userScope);
expect(result.success).toBe(false);
expect(result.command).toBe('provider');
expect(result.message).toBe('Command failed due to an internal error.');
expect(result.message).not.toContain(distinctiveRawFailure);
expect(result.message).not.toContain('ECONNREFUSED');
// The real exception is still logged server-side, as the raw Error
// object itself (not stringified/interpolated into the log message).
expect(loggerErrorSpy).toHaveBeenCalled();
const loggedRawError = loggerErrorSpy.mock.calls.some((call) => call.includes(rawError));
expect(loggedRawError).toBe(true);
loggerErrorSpy.mockRestore();
});
// Inner catch sanitization (P3-5 operator ruling): every catch in
// command-executor.service.ts that returns a SlashCommandResultPayload
// must sanitize the client-facing message the same way the top-level
// catch does, while still logging the raw exception server-side.
it('/agent new sanitizes agent-creation failures, logging the raw exception but never returning it to the client', async () => {
const marker = new Error('distinctive-agent-create-failure-token-A17f');
mockBrain.agents.create.mockRejectedValueOnce(marker);
const loggerErrorSpy = vi.spyOn(Logger.prototype, 'error').mockImplementation(() => undefined);
const payload: SlashCommandPayload = {
command: 'agent',
args: 'new my-new-agent',
conversationId,
};
const result = await service.execute(payload, userScope);
expect(result.success).toBe(false);
expect(result.command).toBe('agent');
expect(result.message).toBe('Failed to create agent due to an internal error.');
expect(result.message).not.toContain('distinctive-agent-create-failure-token-A17f');
expect(loggerErrorSpy).toHaveBeenCalled();
const loggedRawError = loggerErrorSpy.mock.calls.some((call) => call.includes(marker));
expect(loggedRawError).toBe(true);
loggerErrorSpy.mockRestore();
});
it('/agent <name> switch sanitizes agent-lookup failures, logging the raw exception but never returning it to the client', async () => {
const marker = new Error('distinctive-agent-switch-failure-token-B29c');
mockBrain.agents.findByName.mockRejectedValueOnce(marker);
const loggerErrorSpy = vi.spyOn(Logger.prototype, 'error').mockImplementation(() => undefined);
const payload: SlashCommandPayload = {
command: 'agent',
args: 'some-other-agent',
conversationId,
};
const result = await service.execute(payload, userScope);
expect(result.success).toBe(false);
expect(result.command).toBe('agent');
expect(result.message).toBe('Failed to switch agent due to an internal error.');
expect(result.message).not.toContain('distinctive-agent-switch-failure-token-B29c');
expect(loggerErrorSpy).toHaveBeenCalled();
const loggedRawError = loggerErrorSpy.mock.calls.some((call) => call.includes(marker));
expect(loggedRawError).toBe(true);
loggerErrorSpy.mockRestore();
});
it('/mcp reconnect sanitizes MCP client failures, logging the raw exception but never returning it to the client', async () => {
const marker = new Error('distinctive-mcp-reconnect-failure-token-C33e');
const mockMcpClient = {
reconnectServer: vi.fn().mockRejectedValue(marker),
getServerStatuses: vi.fn(() => []),
getToolDefinitions: vi.fn(() => []),
};
const mcpService = buildService(mockRedis, mockMcpClient);
const loggerErrorSpy = vi.spyOn(Logger.prototype, 'error').mockImplementation(() => undefined);
const payload: SlashCommandPayload = {
command: 'mcp',
args: 'reconnect my-server',
conversationId,
};
const result = await mcpService.execute(payload, userScope);
expect(result.success).toBe(false);
expect(result.command).toBe('mcp');
expect(result.message).toBe(
'Failed to reconnect MCP server "my-server" due to an internal error.',
);
expect(result.message).not.toContain('distinctive-mcp-reconnect-failure-token-C33e');
expect(loggerErrorSpy).toHaveBeenCalled();
const loggedRawError = loggerErrorSpy.mock.calls.some((call) => call.includes(marker));
expect(loggedRawError).toBe(true);
loggerErrorSpy.mockRestore();
});
}); });
@@ -36,6 +36,12 @@ const authorization = {
), ),
}; };
const mockMcpClient = {
getServerStatuses: vi.fn(() => []),
getToolDefinitions: vi.fn(() => []),
reconnectServer: vi.fn().mockResolvedValue(undefined),
};
function buildExecutor(authorizationService: unknown = authorization): CommandExecutorService { function buildExecutor(authorizationService: unknown = authorization): CommandExecutorService {
return new CommandExecutorService( return new CommandExecutorService(
registry as never, registry as never,
@@ -46,7 +52,7 @@ function buildExecutor(authorizationService: unknown = authorization): CommandEx
{ agents: {} } as never, { agents: {} } as never,
null, null,
null, null,
null, mockMcpClient as never,
authorizationService as never, authorizationService as never,
); );
} }
@@ -34,9 +34,7 @@ export class CommandExecutorService {
@Optional() @Optional()
@Inject(forwardRef(() => ChatGateway)) @Inject(forwardRef(() => ChatGateway))
private readonly chatGateway: ChatGateway | null, private readonly chatGateway: ChatGateway | null,
@Optional() @Inject(McpClientService) private readonly mcpClient: McpClientService,
@Inject(McpClientService)
private readonly mcpClient: McpClientService | null,
@Optional() @Optional()
@Inject(CommandAuthorizationService) @Inject(CommandAuthorizationService)
private readonly authorization: CommandAuthorizationService | null = null, private readonly authorization: CommandAuthorizationService | null = null,
@@ -159,8 +157,13 @@ export class CommandExecutorService {
}; };
} }
} catch (err) { } catch (err) {
this.logger.error(`Command /${command} failed: ${err}`); this.logger.error(`Command /${command} failed`, err);
return { command, conversationId, success: false, message: String(err) }; return {
command,
conversationId,
success: false,
message: 'Command failed due to an internal error.',
};
} }
} }
@@ -336,11 +339,11 @@ export class CommandExecutorService {
data: { agentId: newAgent.id, agentName: newAgent.name }, data: { agentId: newAgent.id, agentName: newAgent.name },
}; };
} catch (err) { } catch (err) {
this.logger.error(`Failed to create agent: ${err}`); this.logger.error(`Failed to create agent "${namePart}" for user ${userId}`, err);
return { return {
command: 'agent', command: 'agent',
success: false, success: false,
message: `Failed to create agent: ${String(err)}`, message: 'Failed to create agent due to an internal error.',
conversationId, conversationId,
}; };
} }
@@ -391,11 +394,11 @@ export class CommandExecutorService {
data: { agentId: agentConfig.id, agentName: agentConfig.name, model: agentConfig.model }, data: { agentId: agentConfig.id, agentName: agentConfig.name, model: agentConfig.model },
}; };
} catch (err) { } catch (err) {
this.logger.error(`Failed to switch agent "${agentName}": ${err}`); this.logger.error(`Failed to switch agent "${agentName}"`, err);
return { return {
command: 'agent', command: 'agent',
success: false, success: false,
message: `Failed to switch agent: ${String(err)}`, message: 'Failed to switch agent due to an internal error.',
conversationId, conversationId,
}; };
} }
@@ -543,15 +546,6 @@ export class CommandExecutorService {
args: string | null, args: string | null,
conversationId: string, conversationId: string,
): Promise<SlashCommandResultPayload> { ): Promise<SlashCommandResultPayload> {
if (!this.mcpClient) {
return {
command: 'mcp',
conversationId,
success: false,
message: 'MCP client service is not available.',
};
}
const action = args?.trim().split(/\s+/)[0] ?? 'status'; const action = args?.trim().split(/\s+/)[0] ?? 'status';
switch (action) { switch (action) {
@@ -608,11 +602,12 @@ export class CommandExecutorService {
message: `MCP server "${serverName}" reconnected successfully.`, message: `MCP server "${serverName}" reconnected successfully.`,
}; };
} catch (err) { } catch (err) {
this.logger.error(`Failed to reconnect MCP server "${serverName}"`, err);
return { return {
command: 'mcp', command: 'mcp',
conversationId, conversationId,
success: false, success: false,
message: `Failed to reconnect MCP server "${serverName}": ${err instanceof Error ? err.message : String(err)}`, message: `Failed to reconnect MCP server "${serverName}" due to an internal error.`,
}; };
} }
} }
@@ -11,6 +11,8 @@
* - Unknown command returns descriptive error * - Unknown command returns descriptive error
*/ */
import { describe, it, expect, vi, beforeEach } from 'vitest'; import { describe, it, expect, vi, beforeEach } from 'vitest';
import { CommandsModule } from './commands.module.js';
import { McpClientModule } from '../mcp-client/mcp-client.module.js';
import { CommandRegistryService } from './command-registry.service.js'; import { CommandRegistryService } from './command-registry.service.js';
import { CommandExecutorService } from './command-executor.service.js'; import { CommandExecutorService } from './command-executor.service.js';
import type { SlashCommandPayload } from '@mosaicstack/types'; import type { SlashCommandPayload } from '@mosaicstack/types';
@@ -47,6 +49,12 @@ const mockBrain = {
}, },
}; };
const mockMcpClient = {
getServerStatuses: vi.fn(() => []),
getToolDefinitions: vi.fn(() => []),
reconnectServer: vi.fn().mockResolvedValue(undefined),
};
// ─── Helpers ───────────────────────────────────────────────────────────────── // ─── Helpers ─────────────────────────────────────────────────────────────────
function buildRegistry(): CommandRegistryService { function buildRegistry(): CommandRegistryService {
@@ -65,7 +73,7 @@ function buildExecutor(registry: CommandRegistryService): CommandExecutorService
mockBrain as never, mockBrain as never,
null, // reloadService (optional) null, // reloadService (optional)
null, // chatGateway (optional) null, // chatGateway (optional)
null, // mcpClient (optional) mockMcpClient as never,
); );
} }
@@ -153,6 +161,15 @@ describe('CommandRegistryService — integration', () => {
} }
}); });
// ─── Module Wiring Tests ──────────────────────────────────────────────────────
describe('CommandsModule — Nest wiring', () => {
it('CommandsModule imports McpClientModule in its Nest metadata', () => {
const imports = Reflect.getMetadata('imports', CommandsModule) ?? [];
expect(imports).toContain(McpClientModule);
});
});
// ─── Executor Tests ─────────────────────────────────────────────────────────── // ─── Executor Tests ───────────────────────────────────────────────────────────
describe('CommandExecutorService — integration', () => { describe('CommandExecutorService — integration', () => {
@@ -259,4 +276,14 @@ describe('CommandExecutorService — integration', () => {
expect(result.command).toBe(cmd); expect(result.command).toBe(cmd);
}); });
} }
// /mcp status reaches the required McpClientService and never reports it unavailable
it('/mcp status calls the wired McpClientService and reports the no-servers message', async () => {
const payload: SlashCommandPayload = { command: 'mcp', conversationId };
const result = await executor.execute(payload, userScope);
expect(mockMcpClient.getServerStatuses).toHaveBeenCalledOnce();
expect(result.success).toBe(true);
expect(result.message).toContain('No MCP servers configured.');
expect(result.message).not.toBe('MCP client service is not available.');
});
}); });
+7 -1
View File
@@ -4,6 +4,7 @@ import type { MosaicConfig } from '@mosaicstack/config';
import { MOSAIC_CONFIG } from '../config/config.module.js'; import { MOSAIC_CONFIG } from '../config/config.module.js';
import { ChatModule } from '../chat/chat.module.js'; import { ChatModule } from '../chat/chat.module.js';
import { GCModule } from '../gc/gc.module.js'; import { GCModule } from '../gc/gc.module.js';
import { McpClientModule } from '../mcp-client/mcp-client.module.js';
import { ReloadModule } from '../reload/reload.module.js'; import { ReloadModule } from '../reload/reload.module.js';
import { CommandAuthorizationService } from './command-authorization.service.js'; import { CommandAuthorizationService } from './command-authorization.service.js';
import { CommandExecutorService } from './command-executor.service.js'; import { CommandExecutorService } from './command-executor.service.js';
@@ -14,7 +15,12 @@ import { COMMANDS_REDIS } from './commands.tokens.js';
const COMMANDS_QUEUE_HANDLE = 'COMMANDS_QUEUE_HANDLE'; const COMMANDS_QUEUE_HANDLE = 'COMMANDS_QUEUE_HANDLE';
@Module({ @Module({
imports: [GCModule, forwardRef(() => ReloadModule), forwardRef(() => ChatModule)], imports: [
GCModule,
McpClientModule,
forwardRef(() => ReloadModule),
forwardRef(() => ChatModule),
],
providers: [ providers: [
{ {
provide: COMMANDS_QUEUE_HANDLE, provide: COMMANDS_QUEUE_HANDLE,
@@ -0,0 +1,44 @@
import { Logger } from '@nestjs/common';
import { Client } from '@modelcontextprotocol/sdk/client/index.js';
import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
import { McpClientService } from './mcp-client.service.js';
const MCP_LEAK_MARKER = 'MCP_LEAK_MARKER /srv/secret';
describe('McpClientService — failed connect error sanitization', () => {
const originalMcpServers = process.env['MCP_SERVERS'];
beforeEach(() => {
process.env['MCP_SERVERS'] = JSON.stringify([
{ name: 'leaky-server', url: 'http://localhost:9999/mcp' },
]);
});
afterEach(() => {
vi.restoreAllMocks();
if (originalMcpServers === undefined) {
delete process.env['MCP_SERVERS'];
} else {
process.env['MCP_SERVERS'] = originalMcpServers;
}
});
it('stores a generic serverEntry.error while logging the raw exception server-side', async () => {
vi.spyOn(Client.prototype, 'connect').mockRejectedValue(new Error(MCP_LEAK_MARKER));
const errorSpy = vi.spyOn(Logger.prototype, 'error').mockImplementation(() => undefined);
const service = new McpClientService();
await service.onModuleInit();
const statuses = service.getServerStatuses();
expect(statuses).toHaveLength(1);
expect(statuses[0]?.connected).toBe(false);
expect(statuses[0]?.error).toBe('Connection failed (see server logs).');
expect(statuses[0]?.error).not.toContain(MCP_LEAK_MARKER);
const loggedRawMarker = errorSpy.mock.calls.some((call) =>
call.some((arg) => typeof arg === 'string' && arg.includes(MCP_LEAK_MARKER)),
);
expect(loggedRawMarker).toBe(true);
});
});
@@ -189,7 +189,7 @@ export class McpClientService implements OnModuleInit, OnModuleDestroy {
); );
} catch (err) { } catch (err) {
const message = err instanceof Error ? err.message : String(err); const message = err instanceof Error ? err.message : String(err);
serverEntry.error = message; serverEntry.error = 'Connection failed (see server logs).';
serverEntry.connected = false; serverEntry.connected = false;
this.logger.error(`Failed to connect to MCP server "${config.name}": ${message}`); this.logger.error(`Failed to connect to MCP server "${config.name}": ${message}`);
} }
@@ -1,5 +1,8 @@
import { Logger } from '@nestjs/common';
import { describe, expect, it, vi } from 'vitest'; import { describe, expect, it, vi } from 'vitest';
import type { SlashCommandPayload, SystemReloadPayload } from '@mosaicstack/types';
import { ReloadService } from './reload.service.js'; import { ReloadService } from './reload.service.js';
import { CommandExecutorService } from '../commands/command-executor.service.js';
function createMockCommandRegistry() { function createMockCommandRegistry() {
return { return {
@@ -104,3 +107,85 @@ describe('ReloadService', () => {
expect(() => service.registerPlugin('my-plugin', {})).not.toThrow(); expect(() => service.registerPlugin('my-plugin', {})).not.toThrow();
}); });
}); });
describe('ReloadService — /reload command sanitizes plugin errors', () => {
it('generic per-plugin errors reach the chat surface while raw markers stay server-side only', async () => {
const registry = {
getManifest: vi.fn().mockReturnValue({
version: 1,
commands: [
{ name: 'reload', aliases: [], scope: 'core', execution: 'socket', available: true },
],
skills: [],
}),
};
const reloadService = new ReloadService(registry as never);
const RELOAD_LOAD_LEAK_MARKER = 'RELOAD_LOAD_LEAK_MARKER /srv/load-secret';
const RELOAD_UNLOAD_LEAK_MARKER = 'RELOAD_UNLOAD_LEAK_MARKER /srv/unload-secret';
reloadService.registerPlugin('unload-fails', {
pluginName: 'unload-fails',
onLoad: vi.fn().mockResolvedValue(undefined),
onUnload: vi.fn().mockRejectedValue(new Error(RELOAD_UNLOAD_LEAK_MARKER)),
});
reloadService.registerPlugin('load-fails', {
pluginName: 'load-fails',
onLoad: vi.fn().mockRejectedValue(new Error(RELOAD_LOAD_LEAK_MARKER)),
onUnload: vi.fn().mockResolvedValue(undefined),
});
const errorSpy = vi.spyOn(Logger.prototype, 'error').mockImplementation(() => undefined);
const broadcastReload = vi.fn();
const mockChatGateway = { broadcastReload };
const mockAgentService = { getSession: vi.fn(), applyAgentConfig: vi.fn() };
const mockSystemOverride = { set: vi.fn(), get: vi.fn(), clear: vi.fn() };
const mockSessionGC = { sweepOrphans: vi.fn() };
const mockBrain = { agents: { findByName: vi.fn(), findById: vi.fn(), create: vi.fn() } };
const mockMcpClient = {
getServerStatuses: vi.fn(() => []),
getToolDefinitions: vi.fn(() => []),
reconnectServer: vi.fn().mockResolvedValue(undefined),
};
const executor = new CommandExecutorService(
registry as never,
mockAgentService as never,
mockSystemOverride as never,
mockSessionGC as never,
null,
mockBrain as never,
reloadService,
mockChatGateway as never,
mockMcpClient as never,
);
const payload: SlashCommandPayload = { command: 'reload', conversationId: 'conv-1' };
const result = await executor.execute(payload, { userId: 'user-1', tenantId: 'user-1' });
expect(result.success).toBe(true);
expect(result.message).toContain('unload-fails: unload failed (internal error)');
expect(result.message).toContain('load-fails: load failed (internal error)');
expect(result.message).not.toContain(RELOAD_UNLOAD_LEAK_MARKER);
expect(result.message).not.toContain(RELOAD_LOAD_LEAK_MARKER);
expect(broadcastReload).toHaveBeenCalledOnce();
const broadcastPayload = broadcastReload.mock.calls[0]?.[0] as SystemReloadPayload;
expect(broadcastPayload.message).toContain('unload-fails: unload failed (internal error)');
expect(broadcastPayload.message).toContain('load-fails: load failed (internal error)');
expect(broadcastPayload.message).not.toContain(RELOAD_UNLOAD_LEAK_MARKER);
expect(broadcastPayload.message).not.toContain(RELOAD_LOAD_LEAK_MARKER);
const loggedUnloadMarker = errorSpy.mock.calls.some((call) =>
call.some((arg) => typeof arg === 'string' && arg.includes(RELOAD_UNLOAD_LEAK_MARKER)),
);
const loggedLoadMarker = errorSpy.mock.calls.some((call) =>
call.some((arg) => typeof arg === 'string' && arg.includes(RELOAD_LOAD_LEAK_MARKER)),
);
expect(loggedUnloadMarker).toBe(true);
expect(loggedLoadMarker).toBe(true);
errorSpy.mockRestore();
});
});
+4 -2
View File
@@ -58,7 +58,8 @@ export class ReloadService implements OnApplicationBootstrap, OnApplicationShutd
await plugin.onUnload(); await plugin.onUnload();
reloaded.push(name); reloaded.push(name);
} catch (err) { } catch (err) {
errors.push(`${name}: unload failed — ${err}`); this.logger.error(`Plugin "${name}" failed during onUnload: ${err}`);
errors.push(`${name}: unload failed (internal error)`);
} }
} }
} }
@@ -69,7 +70,8 @@ export class ReloadService implements OnApplicationBootstrap, OnApplicationShutd
try { try {
await plugin.onLoad(); await plugin.onLoad();
} catch (err) { } catch (err) {
errors.push(`${name}: load failed ${err}`); this.logger.error(`Plugin "${name}" failed during onLoad: ${err}`);
errors.push(`${name}: load failed (internal error)`);
} }
} }
} }
+3 -2
View File
@@ -5,16 +5,17 @@
"scripts": { "scripts": {
"build": "node ../../scripts/build-web.mjs", "build": "node ../../scripts/build-web.mjs",
"build:vite": "vite build", "build:vite": "vite build",
"dev": "next dev", "dev": "next dev -p 3101",
"dev:vite": "vite", "dev:vite": "vite",
"lint": "eslint src", "lint": "eslint src",
"typecheck": "tsc --noEmit", "typecheck": "tsc --noEmit",
"test": "vitest run --passWithNoTests", "test": "vitest run --passWithNoTests",
"test:e2e": "playwright test", "test:e2e": "playwright test",
"start": "next start" "start": "next start -p 3101"
}, },
"dependencies": { "dependencies": {
"@mosaicstack/design-tokens": "workspace:^", "@mosaicstack/design-tokens": "workspace:^",
"@mosaicstack/types": "workspace:^",
"better-auth": "^1.5.5", "better-auth": "^1.5.5",
"clsx": "^2.1.0", "clsx": "^2.1.0",
"next": "^16.0.0", "next": "^16.0.0",
+61
View File
@@ -0,0 +1,61 @@
// Centralizes the type-only import of the shared `/chat` Socket.IO contract from
// the public `@mosaicstack/types` package. `import type` is erased at compile
// time, so this introduces no runtime dependency — it only reuses the exact
// payload shapes instead of redeclaring them.
import type { Socket } from 'socket.io-client';
import type {
AbortPayload,
AgentEndPayload,
AgentStartPayload,
AgentTextPayload,
AgentThinkingPayload,
ChatMessagePayload,
ClientToServerEvents,
CommandDef,
CommandManifest,
CommandManifestPayload,
ErrorPayload,
MessageAckPayload,
RoutingDecisionInfo,
ServerToClientEvents,
SessionInfoPayload,
SessionUsagePayload,
SetThinkingPayload,
SkillCommandDef,
SlashCommandApprovalResultPayload,
SlashCommandPayload,
SlashCommandResultPayload,
SystemReloadPayload,
ToolEndPayload,
ToolStartPayload,
} from '@mosaicstack/types';
export type {
AbortPayload,
AgentEndPayload,
AgentStartPayload,
AgentTextPayload,
AgentThinkingPayload,
ChatMessagePayload,
ClientToServerEvents,
CommandDef,
CommandManifest,
CommandManifestPayload,
ErrorPayload,
MessageAckPayload,
RoutingDecisionInfo,
ServerToClientEvents,
SessionInfoPayload,
SessionUsagePayload,
SetThinkingPayload,
SkillCommandDef,
SlashCommandApprovalResultPayload,
SlashCommandPayload,
SlashCommandResultPayload,
SystemReloadPayload,
ToolEndPayload,
ToolStartPayload,
};
/** The `/chat` namespace socket, narrowed to the exact typed event contract. */
export type ChatSocket = Socket<ServerToClientEvents, ClientToServerEvents>;
+58 -12
View File
@@ -10,30 +10,52 @@ vi.mock('socket.io-client', () => ({
import { destroySocket, getSocket } from './socket'; import { destroySocket, getSocket } from './socket';
describe('chat socket', () => { interface MockChatSocket {
let disconnectHandler: (() => void) | undefined; on: ReturnType<typeof vi.fn>;
offAny: ReturnType<typeof vi.fn>;
disconnect: ReturnType<typeof vi.fn>;
/** Test-only helper: fires every handler registered for `event` via
* `.on`, mirroring how a real socket.io-client instance invokes its own
* listeners (e.g. calling the registered `disconnect` handler(s) on a
* real transient disconnect). */
trigger(event: string): void;
}
beforeEach(() => { function createMockSocket(): MockChatSocket {
disconnectHandler = undefined; const handlers = new Map<string, Set<() => void>>();
ioMock.mockReset(); const mockSocket: MockChatSocket = {
const mockSocket = {
on: vi.fn((event: string, handler: () => void) => { on: vi.fn((event: string, handler: () => void) => {
if (event === 'disconnect') disconnectHandler = handler; if (!handlers.has(event)) handlers.set(event, new Set());
handlers.get(event)?.add(handler);
return mockSocket; return mockSocket;
}), }),
offAny: vi.fn(() => mockSocket), offAny: vi.fn(() => mockSocket),
disconnect: vi.fn(() => mockSocket), disconnect: vi.fn(() => mockSocket),
trigger(event: string): void {
for (const handler of handlers.get(event) ?? []) handler();
},
}; };
return mockSocket;
}
ioMock.mockReturnValue(mockSocket); let currentMock!: MockChatSocket;
describe('chat socket', () => {
beforeEach(() => {
ioMock.mockReset();
// A fresh object per io() call so identity assertions (same singleton vs.
// a genuinely new instance) are meaningful.
ioMock.mockImplementation(() => {
currentMock = createMockSocket();
return currentMock;
});
}); });
afterEach(() => { afterEach(() => {
destroySocket(); destroySocket();
}); });
it('creates one same-origin /chat namespace socket until it disconnects', () => { it('creates one same-origin /chat namespace socket', () => {
const first = getSocket(); const first = getSocket();
const second = getSocket(); const second = getSocket();
@@ -44,9 +66,33 @@ describe('chat socket', () => {
autoConnect: false, autoConnect: false,
transports: ['websocket', 'polling'], transports: ['websocket', 'polling'],
}); });
});
disconnectHandler?.(); it('keeps the same singleton instance across a transient disconnect', () => {
getSocket(); const first = getSocket();
// socket.ts must not react to a real socket's `disconnect` event by
// nulling the singleton — it registers no such handler at all now.
// Actually fire every handler registered via `.on('disconnect', ...)`
// (mirroring a real socket.io-client reconnect) instead of merely
// calling getSocket() again: this is what makes the test fail if
// production reintroduces `socket.on('disconnect', () => { socket =
// null; })`, since that handler would run here and null the singleton
// before the next getSocket() call.
currentMock.trigger('disconnect');
const second = getSocket();
expect(second).toBe(first);
expect(ioMock).toHaveBeenCalledOnce();
});
it('only creates a new singleton after an explicit destroySocket()', () => {
const first = getSocket();
destroySocket();
const second = getSocket();
expect(second).not.toBe(first);
expect(ioMock).toHaveBeenCalledTimes(2); expect(ioMock).toHaveBeenCalledTimes(2);
}); });
}); });
+16 -10
View File
@@ -1,21 +1,27 @@
import { io, type Socket } from 'socket.io-client'; import { io } from 'socket.io-client';
import type { ChatSocket } from './chat-contract';
let socket: Socket | null = null; let socket: ChatSocket | null = null;
export function getSocket(): Socket { export function getSocket(): ChatSocket {
if (!socket) { if (!socket) {
// socket.io-client 4.8.3's `io()` factory declaration always returns the
// default unparameterized Socket (it accepts no <ListenEvents, EmitEvents>
// generics), so this one cast is the unavoidable boundary between that and the
// typed `/chat` contract. Every other call site uses the resulting ChatSocket
// with no further assertions.
socket = io('/chat', { socket = io('/chat', {
withCredentials: true, withCredentials: true,
autoConnect: false, autoConnect: false,
transports: ['websocket', 'polling'], transports: ['websocket', 'polling'],
}); }) as unknown as ChatSocket;
// Reset singleton reference when socket is fully closed so the next // A transient `disconnect` (network blip, server restart) must NOT null
// getSocket() call creates a fresh instance instead of returning a // the singleton: socket.io-client auto-reconnects this same instance,
// closed/dead socket. // and its listeners stay registered across that reconnect. Nulling here
socket.on('disconnect', () => { // previously orphaned those listeners on the next getSocket() call by
socket = null; // handing back a brand-new, unconnected instance. Only destroySocket()
}); // (an explicit, intentional teardown) may reset the singleton.
} }
return socket; return socket;
} }
+22 -4
View File
@@ -3,6 +3,16 @@ import { createBrowserRouter, Navigate, Outlet, type RouteObject } from 'react-r
import { LoginPage } from '@/spa/pages/login'; import { LoginPage } from '@/spa/pages/login';
import { RegisterPage } from '@/spa/pages/register'; import { RegisterPage } from '@/spa/pages/register';
import { SsoCallbackPage } from '@/spa/pages/sso-callback'; import { SsoCallbackPage } from '@/spa/pages/sso-callback';
import { ChatPage } from '@/spa/pages/chat';
import { ChatRouteErrorBoundary } from '@/spa/pages/chat-error-boundary';
import { ProjectDetailPage } from '@/spa/pages/project-detail';
import { ProjectsPage } from '@/spa/pages/projects';
import {
ProjectDetailRouteErrorBoundary,
ProjectsRouteErrorBoundary,
TasksRouteErrorBoundary,
} from '@/spa/pages/resource-route-error-boundaries';
import { TasksPage } from '@/spa/pages/tasks';
import { AuthGuard, GuestGuard } from '@/spa/guards'; import { AuthGuard, GuestGuard } from '@/spa/guards';
import { Placeholder } from '@/spa/placeholder'; import { Placeholder } from '@/spa/placeholder';
@@ -34,10 +44,18 @@ export const routes: RouteObject[] = [
element: <AuthGuard />, element: <AuthGuard />,
children: [ children: [
{ path: '/', element: <Navigate to="/chat" replace /> }, { path: '/', element: <Navigate to="/chat" replace /> },
{ path: '/chat', element: <Placeholder title="Chat" /> }, { path: '/chat', element: <ChatPage />, errorElement: <ChatRouteErrorBoundary /> },
{ path: '/projects', element: <Placeholder title="Projects" /> }, {
{ path: '/projects/:id', element: <Placeholder title="Project" /> }, path: '/projects',
{ path: '/tasks', element: <Placeholder title="Tasks" /> }, element: <ProjectsPage />,
errorElement: <ProjectsRouteErrorBoundary />,
},
{
path: '/projects/:id',
element: <ProjectDetailPage />,
errorElement: <ProjectDetailRouteErrorBoundary />,
},
{ path: '/tasks', element: <TasksPage />, errorElement: <TasksRouteErrorBoundary /> },
{ path: '/settings', element: <Placeholder title="Settings" /> }, { path: '/settings', element: <Placeholder title="Settings" /> },
{ path: '/admin', element: <Placeholder title="Admin" /> }, { path: '/admin', element: <Placeholder title="Admin" /> },
], ],
@@ -0,0 +1,280 @@
import { act } from 'react';
import { createRoot, type Root } from 'react-dom/client';
import { afterAll, afterEach, beforeAll, describe, expect, it, vi } from 'vitest';
import { CommandsPanel } from './commands-panel';
beforeAll(() => {
Object.defineProperty(globalThis, 'IS_REACT_ACT_ENVIRONMENT', {
configurable: true,
value: true,
});
});
afterAll(() => {
Reflect.deleteProperty(globalThis, 'IS_REACT_ACT_ENVIRONMENT');
});
let root: Root | null;
let container: HTMLElement | null;
async function render(node: Parameters<Root['render']>[0]): Promise<void> {
container = document.createElement('div');
document.body.append(container);
root = createRoot(container);
await act(async () => {
root?.render(node);
});
}
afterEach(async () => {
await act(async () => {
root?.unmount();
});
document.body.replaceChildren();
root = null;
container = null;
});
describe('CommandsPanel', () => {
it('shows the frozen local pendingApproval args in the confirmation area, regardless of misleading server message text', async () => {
await render(
<CommandsPanel
manifest={null}
results={[]}
approval={{
conversationId: 'c1',
command: 'deploy', // matches pendingApproval — this is a legitimately approved request
success: true,
approvalId: 'ap1',
expiresAt: '2026-01-01T00:00:00.000Z',
// Free-text server message claims a different, less alarming target
// than what will actually be sent — the UI must not rely on this.
message: 'This will only affect the staging environment.',
}}
pendingApproval={{ command: 'deploy', args: 'prod' }}
hasConversation
onExecute={vi.fn()}
onApprove={vi.fn()}
onRunApproved={vi.fn()}
/>,
);
// The exact frozen combined action is visible...
expect(container?.textContent).toContain('/deploy');
expect(container?.textContent).toContain('prod');
// ...and the misleading server free-text is never shown next to it.
expect(container?.textContent).not.toContain('staging environment');
});
it('does not throw when a manifest commands entry is null', async () => {
const manifest = {
commands: [
null,
{
name: 'model',
aliases: [],
description: 'Change the active model',
scope: 'core',
execution: 'socket',
available: true,
},
],
skills: [null],
version: 1,
} as unknown as Parameters<typeof CommandsPanel>[0]['manifest'];
await expect(
render(
<CommandsPanel
manifest={manifest}
results={[]}
approval={null}
pendingApproval={null}
hasConversation={false}
onExecute={vi.fn()}
onApprove={vi.fn()}
onRunApproved={vi.fn()}
/>,
),
).resolves.not.toThrow();
expect(container?.textContent).toContain('model');
});
it('shows an explicit no-args fallback when the frozen pendingApproval has no args', async () => {
await render(
<CommandsPanel
manifest={null}
results={[]}
approval={{
conversationId: 'c1',
command: 'deploy',
success: true,
approvalId: 'ap1',
expiresAt: '2026-01-01T00:00:00.000Z',
}}
pendingApproval={{ command: 'deploy' }}
hasConversation
onExecute={vi.fn()}
onApprove={vi.fn()}
onRunApproved={vi.fn()}
/>,
);
expect(container?.textContent?.toLowerCase()).toContain('no args');
});
it('renders skills from a skills-only manifest', async () => {
await render(
<CommandsPanel
manifest={{
commands: [],
skills: [{ name: 'brave-search', description: 'Search the web', available: true }],
version: 1,
}}
results={[]}
approval={null}
pendingApproval={null}
hasConversation={false}
onExecute={vi.fn()}
onApprove={vi.fn()}
onRunApproved={vi.fn()}
/>,
);
expect(container?.textContent).toContain('brave-search');
expect(container?.textContent).toContain('Search the web');
});
it('does not show the Run affordance when approval.success/approvalId are objects, even though command matches pendingApproval', async () => {
const approval = {
conversationId: 'c1',
command: 'deploy',
success: { truthy: 'object' },
approvalId: { also: 'object' },
} as unknown as Parameters<typeof CommandsPanel>[0]['approval'];
await render(
<CommandsPanel
manifest={null}
results={[]}
approval={approval}
pendingApproval={{ command: 'deploy', args: 'prod' }}
hasConversation
onExecute={vi.fn()}
onApprove={vi.fn()}
onRunApproved={vi.fn()}
/>,
);
expect(
[...(container?.querySelectorAll('button') ?? [])].some((button) =>
button.textContent?.includes('Run approved command'),
),
).toBe(false);
});
it('shows the guarded server-provided denial reason for a denied approval', async () => {
await render(
<CommandsPanel
manifest={null}
results={[]}
approval={{
conversationId: 'c1',
command: 'deploy',
success: false,
message: 'Not authorized',
}}
pendingApproval={{ command: 'deploy', args: 'prod' }}
hasConversation
onExecute={vi.fn()}
onApprove={vi.fn()}
onRunApproved={vi.fn()}
/>,
);
expect(container?.textContent).toContain('Not authorized');
});
it('falls back to a stable "Denied." copy when a denial has no usable message', async () => {
await render(
<CommandsPanel
manifest={null}
results={[]}
approval={{ conversationId: 'c1', command: 'deploy', success: false }}
pendingApproval={{ command: 'deploy', args: 'prod' }}
hasConversation
onExecute={vi.fn()}
onApprove={vi.fn()}
onRunApproved={vi.fn()}
/>,
);
expect(container?.textContent).toContain('Denied.');
});
it('shows the guarded contract-provided reason for a failed command result, falling back to a stable copy only when absent', async () => {
await render(
<CommandsPanel
manifest={null}
results={[
{ conversationId: 'c1', command: 'model', success: false, message: 'Unknown model' },
{ conversationId: 'c1', command: 'deploy', success: false },
]}
approval={null}
pendingApproval={null}
hasConversation={false}
onExecute={vi.fn()}
onApprove={vi.fn()}
onRunApproved={vi.fn()}
/>,
);
expect(container?.textContent).toContain('Unknown model');
expect(container?.textContent).toContain('Command failed.');
});
it('bounds an oversized command result message at the render site as defense-in-depth', async () => {
const hostileMessage = 'y'.repeat(50_000);
await render(
<CommandsPanel
manifest={null}
results={[
{ conversationId: 'c1', command: 'model', success: false, message: hostileMessage },
]}
approval={null}
pendingApproval={null}
hasConversation={false}
onExecute={vi.fn()}
onApprove={vi.fn()}
onRunApproved={vi.fn()}
/>,
);
const text = container?.textContent ?? '';
expect(text.length).toBeLessThan(hostileMessage.length);
});
it('does not throw when the manifest fields are malformed (non-array commands/skills)', async () => {
const manifest = {
commands: 'not-an-array',
skills: null,
version: 1,
} as unknown as Parameters<typeof CommandsPanel>[0]['manifest'];
await expect(
render(
<CommandsPanel
manifest={manifest}
results={[]}
approval={null}
pendingApproval={null}
hasConversation={false}
onExecute={vi.fn()}
onApprove={vi.fn()}
onRunApproved={vi.fn()}
/>,
),
).resolves.not.toThrow();
});
});
+164
View File
@@ -0,0 +1,164 @@
import { useState, type ReactElement } from 'react';
import type { PendingApproval } from './use-chat-connection';
import { MAX_COMMAND_MESSAGE_CHARS } from './limits';
import { asNonEmptyString, asString } from './runtime-guards';
import type {
CommandManifest,
SlashCommandApprovalResultPayload,
SlashCommandResultPayload,
} from '@/lib/chat-contract';
/** Stable fallback copy shown for a failed command only when the server's
* own guarded, non-empty `message` (e.g. "Unknown model") is absent or
* malformed — the structured contract reason itself is otherwise shown
* directly, never a raw thrown exception, stack trace, or object value. */
const COMMAND_FAILURE_COPY = 'Command failed.';
/** Render-site defense-in-depth: `use-chat-connection.ts` already bounds a
* stored command:result message at ingestion, but this component must never
* assume every caller went through that path — bounding again here means a
* hostile/oversized message can never force an unbounded render. */
function boundMessage(value: string): string {
return value.length > MAX_COMMAND_MESSAGE_CHARS
? value.slice(0, MAX_COMMAND_MESSAGE_CHARS)
: value;
}
interface CommandsPanelProps {
manifest: CommandManifest | null;
results: SlashCommandResultPayload[];
approval: SlashCommandApprovalResultPayload | null;
pendingApproval: PendingApproval | null;
hasConversation: boolean;
onExecute: (input: { command: string; args?: string }) => void;
onApprove: (input: { command: string; args?: string }) => void;
onRunApproved: () => void;
}
export function CommandsPanel({
manifest,
results,
approval,
pendingApproval,
hasConversation,
onExecute,
onApprove,
onRunApproved,
}: CommandsPanelProps): ReactElement {
const [command, setCommand] = useState('');
const [args, setArgs] = useState('');
// Defense-in-depth: the reducer already normalizes success/approvalId
// before storing `approval`, but a matching command string alone must
// never be trusted here either — require the literal boolean `true` and a
// non-empty string approvalId, not merely truthy values.
const canRunApproved =
approval?.success === true &&
typeof approval.approvalId === 'string' &&
approval.approvalId.length > 0 &&
!!pendingApproval &&
pendingApproval.command === approval.command;
// A manifest arrives from the server as untyped JSON at runtime — guard
// both collections before mapping so a malformed manifest cannot throw.
const commands = Array.isArray(manifest?.commands) ? manifest.commands : [];
const skills = Array.isArray(manifest?.skills) ? manifest.skills : [];
return (
<section aria-label="Commands" className="flex flex-col gap-2 border-b px-4 py-3 text-xs">
{commands.length > 0 ? (
<ul aria-label="Available commands" className="flex flex-col gap-1">
{commands.map((cmd, index) => (
<li key={asString(cmd?.name) || `cmd-${index}`}>
<strong>/{asString(cmd?.name)}</strong> {asString(cmd?.description)}
</li>
))}
</ul>
) : null}
{skills.length > 0 ? (
<ul aria-label="Available skills" className="flex flex-col gap-1">
{skills.map((skill, index) => (
<li key={asString(skill?.name) || `skill-${index}`}>
<strong>/skill:{asString(skill?.name)}</strong> {asString(skill?.description)}
</li>
))}
</ul>
) : null}
<div className="flex flex-wrap items-center gap-2">
<input
aria-label="Command name"
value={command}
onChange={(event) => setCommand(event.target.value)}
placeholder="command"
/>
<input
aria-label="Command arguments"
value={args}
onChange={(event) => setArgs(event.target.value)}
placeholder="args (optional)"
/>
<button
type="button"
disabled={!hasConversation || !command.trim()}
onClick={() => onExecute({ command: command.trim(), args: args.trim() || undefined })}
>
Run command
</button>
<button
type="button"
disabled={!hasConversation || !command.trim()}
onClick={() => onApprove({ command: command.trim(), args: args.trim() || undefined })}
>
Request approval
</button>
</div>
{approval ? (
<div role={approval.success ? 'status' : 'alert'} className="flex items-center gap-2">
{/* A successful approval shows stable client copy only — never
the server-controlled approval.message or echoed
approval.command as the primary confirmation. The frozen local
pendingApproval below (not this line) is the sole authoritative
statement of what will run. A denial, by contrast, is not an
execution authority and safely surfaces the guarded structured
reason the server gave (e.g. "Not authorized"), falling back to
a stable copy only when absent/malformed. */}
<span>
{approval.success ? 'Approved.' : asNonEmptyString(approval.message, 'Denied.')}
</span>
{canRunApproved && pendingApproval ? (
<>
{/* Authoritative frozen local command+args — what the click below
will actually emit. The server's `approval` above is display-only
and must never be trusted to represent the executed payload. */}
<span>
Will run: /{pendingApproval.command}{' '}
{pendingApproval.args ? pendingApproval.args : '(no args)'}
</span>
<button type="button" onClick={onRunApproved}>
Run approved command
</button>
</>
) : null}
</div>
) : null}
{results.length > 0 ? (
<ul aria-label="Command results" className="flex flex-col gap-1">
{results.map((result, index) => (
<li key={`${result.command}-${index}`} role={result.success ? 'status' : 'alert'}>
/{asString(result.command)}: {result.success ? 'success' : 'failed'}
{result.success
? typeof result.message === 'string' && result.message
? `${boundMessage(result.message)}`
: ''
: `${boundMessage(asNonEmptyString(result.message, COMMAND_FAILURE_COPY))}`}
</li>
))}
</ul>
) : null}
</section>
);
}
+98
View File
@@ -0,0 +1,98 @@
import { useState, type KeyboardEvent, type ReactElement } from 'react';
interface ComposerProps {
onSend: (input: { content: string; provider?: string; modelId?: string }) => void;
onStop: () => void;
streaming: boolean;
/** True from local send time through server turn startup/ack and
* throughout streaming — a superset of `streaming` that also covers the
* pre-ack window where a second send could otherwise slip through. */
sending: boolean;
hasConversation: boolean;
}
export function Composer({
onSend,
onStop,
streaming,
sending,
hasConversation,
}: ComposerProps): ReactElement {
const [content, setContent] = useState('');
const [provider, setProvider] = useState('');
const [modelId, setModelId] = useState('');
const busy = streaming || sending;
function submit(): void {
if (busy) return;
const trimmed = content.trim();
if (!trimmed) return;
onSend({
content: trimmed,
provider: provider.trim() || undefined,
modelId: modelId.trim() || undefined,
});
setContent('');
}
function handleKeyDown(event: KeyboardEvent<HTMLTextAreaElement>): void {
if (event.key === 'Enter' && !event.shiftKey) {
event.preventDefault();
submit();
}
}
return (
<form
onSubmit={(event) => {
event.preventDefault();
submit();
}}
className="flex flex-col gap-2 border-t p-4"
>
<div className="flex flex-wrap gap-2">
<input
aria-label="Provider"
value={provider}
onChange={(event) => setProvider(event.target.value)}
placeholder="Provider (optional)"
className="rounded border px-2 py-1 text-xs"
/>
<input
aria-label="Model"
value={modelId}
onChange={(event) => setModelId(event.target.value)}
placeholder="Model (optional)"
className="rounded border px-2 py-1 text-xs"
/>
</div>
<div className="flex items-end gap-2">
<textarea
aria-label="Message"
value={content}
onChange={(event) => setContent(event.target.value)}
onKeyDown={handleKeyDown}
rows={2}
placeholder="Message… (Enter to send, Shift+Enter for a new line)"
className="flex-1 resize-none rounded border px-3 py-2 text-sm"
/>
<button
type="submit"
disabled={!content.trim() || busy}
className="rounded px-3 py-2 text-sm font-medium"
>
Send
</button>
<button
type="button"
aria-label="Stop"
disabled={!hasConversation || !streaming}
onClick={onStop}
className="rounded px-3 py-2 text-sm font-medium"
>
Stop
</button>
</div>
</form>
);
}
+22
View File
@@ -0,0 +1,22 @@
/**
* Bounds on server-fed chat state. A hostile or malfunctioning gateway can
* flood any of these collections; caps keep memory/render cost flat instead
* of growing unboundedly for the lifetime of the connection.
*/
/** Max characters retained for the in-flight streamed text/thinking buffers. */
export const MAX_STREAM_CHARS = 20_000;
/** Max transcript turns retained (oldest dropped first). */
export const MAX_MESSAGES = 500;
/** Max tool-call entries (including anomaly entries) retained per turn history. */
export const MAX_TOOLS = 200;
/** Max slash-command results retained. */
export const MAX_COMMAND_RESULTS = 200;
/** Max commands/skills accepted from a single manifest push. */
export const MAX_MANIFEST_ITEMS = 500;
/** Max executed approval IDs remembered for single-flight dedup. */
export const MAX_EXECUTED_APPROVAL_IDS = 200;
/** Max characters retained for a single command:result message — a hostile
* or malfunctioning gateway must not be able to push an unbounded curated
* success/failure reason into state (or, defensively, onto the page). */
export const MAX_COMMAND_MESSAGE_CHARS = 1_000;
@@ -0,0 +1,39 @@
import type { ReactElement } from 'react';
import type { ChatTranscriptMessage } from './use-chat-connection';
interface MessageTranscriptProps {
messages: ChatTranscriptMessage[];
streaming: boolean;
text: string;
}
export function MessageTranscript({
messages,
streaming,
text,
}: MessageTranscriptProps): ReactElement {
return (
<div
role="log"
aria-live="polite"
aria-label="Conversation"
className="flex flex-1 flex-col gap-3 overflow-y-auto p-4"
>
{messages.map((message) => (
<div key={message.id} data-role={message.role} className="whitespace-pre-wrap text-sm">
<span className="font-medium">{message.role === 'user' ? 'You' : 'Assistant'}: </span>
<span>{message.text}</span>
{message.thinking ? (
<div className="pt-1 text-xs italic opacity-70">{message.thinking}</div>
) : null}
</div>
))}
{streaming ? (
<div data-role="assistant-streaming" className="whitespace-pre-wrap text-sm">
<span className="font-medium">Assistant: </span>
<span>{text || 'Thinking…'}</span>
</div>
) : null}
</div>
);
}
+49
View File
@@ -0,0 +1,49 @@
/**
* Socket.IO payloads are only statically typed at the call site — a
* misbehaving or compromised gateway can send anything at runtime. These
* guards protect the dereference sites that would otherwise throw (`.map` on
* a non-array, `.toFixed` on a non-number) or render an object as a React
* child.
*/
export function asString(value: unknown, fallback = ''): string {
return typeof value === 'string' ? value : fallback;
}
/** Like `asString`, but an empty string also falls back — used for guarded
* contract-provided reason strings (e.g. a denial or failure message) where
* an empty string is not a meaningful value to display in place of the
* stable fallback copy. */
export function asNonEmptyString(value: unknown, fallback: string): string {
return typeof value === 'string' && value.length > 0 ? value : fallback;
}
export function asFiniteNumber(value: unknown, fallback = 0): number {
return typeof value === 'number' && Number.isFinite(value) ? value : fallback;
}
/** Like `asFiniteNumber`, but returns `null` on failure instead of a numeric
* fallback — callers that must not fabricate a plausible-looking value (e.g.
* `0 tokens` / `$0.0000` for genuinely unknown usage) use this to render an
* honest "unavailable" label instead. */
export function asFiniteNumberOrNull(value: unknown): number | null {
return typeof value === 'number' && Number.isFinite(value) ? value : null;
}
export function asStringArray(value: unknown): string[] {
return Array.isArray(value) && value.every((item) => typeof item === 'string') ? value : [];
}
export function isRecord(value: unknown): value is Record<string, unknown> {
return typeof value === 'object' && value !== null;
}
/** The single point of truth for what counts as a valid conversation ID
* anywhere a scoped server event may adopt one into state — a non-empty
* string, nothing else. Every site that establishes or compares
* `state.conversationId` against a raw socket payload must route through
* this guard so a malformed first frame (null/object/number/empty string)
* can never be adopted verbatim. */
export function asConversationId(value: unknown): string | null {
return typeof value === 'string' && value.length > 0 ? value : null;
}
+68
View File
@@ -0,0 +1,68 @@
import type { ReactElement } from 'react';
import type { SessionInfoPayload } from '@/lib/chat-contract';
import { MAX_MANIFEST_ITEMS } from './limits';
import { asString, asStringArray } from './runtime-guards';
interface SessionPanelProps {
sessionInfo: SessionInfoPayload | null;
onSetThinking: (level: string) => void;
}
const THINKING_LEVEL_UNAVAILABLE = '';
export function SessionPanel({
sessionInfo,
onSetThinking,
}: SessionPanelProps): ReactElement | null {
if (!sessionInfo) return null;
// The reducer already caps this before storing it, but the render site
// defends independently — a hostile payload must never be able to force
// this <select> to lay out an unbounded number of options.
const availableThinkingLevels = asStringArray(sessionInfo.availableThinkingLevels).slice(
0,
MAX_MANIFEST_ITEMS,
);
const hasThinkingLevels = availableThinkingLevels.length > 0;
return (
<section
aria-label="Session info"
className="flex flex-wrap items-center gap-3 border-b px-4 py-2 text-xs"
>
<span>{asString(sessionInfo.provider, 'unknown')}</span>
<span>{asString(sessionInfo.modelId, 'unknown')}</span>
<label className="flex items-center gap-2">
<span>Thinking level</span>
<select
aria-label="Thinking level"
value={
hasThinkingLevels ? asString(sessionInfo.thinkingLevel) : THINKING_LEVEL_UNAVAILABLE
}
onChange={(event) => {
// The placeholder option is not a real, settable level — a
// malformed availableThinkingLevels list must never let the
// client emit set:thinking for it.
if (!hasThinkingLevels) return;
onSetThinking(event.target.value);
}}
>
{hasThinkingLevels ? (
availableThinkingLevels.map((level) => (
<option key={level} value={level}>
{level}
</option>
))
) : (
<option value={THINKING_LEVEL_UNAVAILABLE}>Thinking level unavailable</option>
)}
</select>
</label>
{sessionInfo.routingDecision ? (
<span title={asString(sessionInfo.routingDecision.ruleName)}>
{asString(sessionInfo.routingDecision.reason)}
</span>
) : null}
</section>
);
}
@@ -0,0 +1,125 @@
import { vi } from 'vitest';
import type { ClientToServerEvents, ServerToClientEvents } from '@/lib/chat-contract';
type ServerEvent = keyof ServerToClientEvents;
type ClientEvent = keyof ClientToServerEvents;
type ServerHandler<K extends ServerEvent> = ServerToClientEvents[K];
type ClientPayload<K extends ClientEvent> = Parameters<ClientToServerEvents[K]>[0];
export interface EmittedEvent<K extends ClientEvent = ClientEvent> {
event: K;
payload: ClientPayload<K>;
}
/** The subset of a Socket.IO `ChatSocket` that `useChatConnection` drives. */
export interface FakeChatSocket {
connected: boolean;
connect(): FakeChatSocket;
on<K extends ServerEvent>(event: K, handler: ServerHandler<K>): FakeChatSocket;
off<K extends ServerEvent>(event: K, handler: ServerHandler<K>): FakeChatSocket;
emit<K extends ClientEvent>(event: K, payload: ClientPayload<K>): FakeChatSocket;
}
/**
* A typed in-memory stand-in for `getSocket()`. Unlike a bare
* `(event: string, payload: unknown) => void` mock, every public method here is
* checked against the real `/chat` contract — a typo'd event name or a payload
* missing a required field fails to compile instead of silently no-op'ing at
* runtime.
*/
/** Socket.IO's built-in connection-state events. Not part of the app-level
* ServerToClientEvents contract, but real sockets always support them and
* `useChatConnection` registers a `disconnect` handler on the real socket. */
type LifecycleEvent = 'connect' | 'disconnect';
export function createFakeChatSocket(): {
socket: FakeChatSocket;
listeners: Map<ServerEvent, Set<(payload: never) => void>>;
emitted: EmittedEvent[];
serverEmit<K extends ServerEvent>(
event: K,
payload: Parameters<ServerToClientEvents[K]>[0],
): void;
/** Escape hatch for malformed-payload tests: bypasses the compile-time
* payload contract to simulate a genuinely untrusted runtime value from the
* server, e.g. a `session:info` with a non-array `availableThinkingLevels`. */
serverEmitRaw(event: ServerEvent, payload: unknown): void;
/** Simulates a transient Socket.IO `disconnect` — fires any handler(s)
* registered via `socket.on('disconnect', ...)` without clearing any
* listeners, mirroring how a real reconnecting socket behaves. */
simulateDisconnect(): void;
/** Simulates socket.io-client's automatic reconnect of the *same*
* instance after a transient disconnect: marks the socket connected again
* and fires any handler(s) registered via `socket.on('connect', ...)`,
* without clearing or replacing any listeners. */
simulateReconnect(): void;
} {
const listeners = new Map<ServerEvent, Set<(payload: never) => void>>();
const emitted: EmittedEvent[] = [];
// Internal storage is intentionally keyed loosely (the per-event handler shape
// varies by K, which a single Map can't express); the generic signatures on the
// exported `socket`/`serverEmit` above and below are what keep test call sites
// type-checked against ServerToClientEvents/ClientToServerEvents.
const socket = {
connected: false,
connect: vi.fn(function connect(this: void) {
socket.connected = true;
return socket;
}),
on: vi.fn(function on(this: void, event: ServerEvent, handler: (payload: never) => void) {
if (!listeners.has(event)) listeners.set(event, new Set());
listeners.get(event)?.add(handler);
return socket;
}),
off: vi.fn(function off(this: void, event: ServerEvent, handler: (payload: never) => void) {
listeners.get(event)?.delete(handler);
return socket;
}),
emit: vi.fn(function emit(this: void, event: ClientEvent, payload: unknown) {
emitted.push({ event, payload } as EmittedEvent);
return socket;
}),
} as unknown as FakeChatSocket;
function serverEmit<K extends ServerEvent>(
event: K,
payload: Parameters<ServerToClientEvents[K]>[0],
): void {
for (const handler of listeners.get(event) ?? []) {
(handler as (payload: Parameters<ServerToClientEvents[K]>[0]) => void)(payload);
}
}
function serverEmitRaw(event: ServerEvent, payload: unknown): void {
for (const handler of listeners.get(event) ?? []) {
(handler as (payload: unknown) => void)(payload);
}
}
function simulateDisconnect(): void {
socket.connected = false;
const lifecycleKey = 'disconnect' satisfies LifecycleEvent as unknown as ServerEvent;
for (const handler of listeners.get(lifecycleKey) ?? []) {
(handler as () => void)();
}
}
function simulateReconnect(): void {
socket.connected = true;
const lifecycleKey = 'connect' satisfies LifecycleEvent as unknown as ServerEvent;
for (const handler of listeners.get(lifecycleKey) ?? []) {
(handler as () => void)();
}
}
return {
socket,
listeners,
emitted,
serverEmit,
serverEmitRaw,
simulateDisconnect,
simulateReconnect,
};
}
@@ -0,0 +1,63 @@
import { act } from 'react';
import { createRoot, type Root } from 'react-dom/client';
import { afterAll, afterEach, beforeAll, describe, expect, it, vi } from 'vitest';
import { ToolCallList } from './tool-call-list';
beforeAll(() => {
Object.defineProperty(globalThis, 'IS_REACT_ACT_ENVIRONMENT', {
configurable: true,
value: true,
});
});
afterAll(() => {
Reflect.deleteProperty(globalThis, 'IS_REACT_ACT_ENVIRONMENT');
});
let root: Root | null;
let container: HTMLElement | null;
async function render(node: Parameters<Root['render']>[0]): Promise<void> {
container = document.createElement('div');
document.body.append(container);
root = createRoot(container);
await act(async () => {
root?.render(node);
});
}
afterEach(async () => {
await act(async () => {
root?.unmount();
});
document.body.replaceChildren();
root = null;
container = null;
});
describe('ToolCallList', () => {
it('renders two entries independently, without a duplicate-key warning, when a valid toolCallId is shared', async () => {
const consoleError = vi.spyOn(console, 'error').mockImplementation(() => {});
await render(
<ToolCallList
tools={[
{ toolCallId: 'dup', toolName: 'search', status: 'success' },
{ toolCallId: 'dup', toolName: 'search', status: 'running' },
]}
/>,
);
const items = [...(container?.querySelectorAll('li') ?? [])];
expect(items).toHaveLength(2);
expect(items[0]?.textContent).toContain('success');
expect(items[1]?.textContent).toContain('running');
const duplicateKeyWarning = consoleError.mock.calls.some((args) =>
args.some((arg) => typeof arg === 'string' && arg.includes('same key')),
);
expect(duplicateKeyWarning).toBe(false);
consoleError.mockRestore();
});
});
+24
View File
@@ -0,0 +1,24 @@
import type { ReactElement } from 'react';
import type { ToolCallState } from './use-chat-connection';
export function ToolCallList({ tools }: { tools: ToolCallState[] }): ReactElement | null {
if (tools.length === 0) return null;
return (
<ul aria-label="Tool calls" className="flex flex-col gap-1 px-4 pb-2 text-xs">
{tools.map((tool, index) => (
<li
// A valid server-controlled toolCallId can legitimately repeat
// (e.g. two tool:start events sharing one id) — keying on it alone
// would give React two identical keys. Pairing it with its
// (stable, append-only) render index keeps every key unique.
key={`${tool.toolCallId}-${index}`}
role={tool.status === 'error' || tool.status === 'anomaly' ? 'alert' : 'status'}
>
{tool.toolName} {' '}
{tool.status === 'anomaly' ? 'unexpected end (unknown tool call)' : tool.status}
</li>
))}
</ul>
);
}
File diff suppressed because it is too large Load Diff
File diff suppressed because it is too large Load Diff
@@ -0,0 +1,87 @@
import { act } from 'react';
import { createRoot, type Root } from 'react-dom/client';
import { createMemoryRouter, RouterProvider, type RouteObject } from 'react-router-dom';
import { afterAll, afterEach, beforeAll, describe, expect, it, vi } from 'vitest';
const { useSessionMock } = vi.hoisted(() => ({
useSessionMock: vi.fn(),
}));
vi.mock('@/lib/auth-client', () => ({
useSession: useSessionMock,
}));
import { routes } from '@/routes';
beforeAll(() => {
Object.defineProperty(globalThis, 'IS_REACT_ACT_ENVIRONMENT', {
configurable: true,
value: true,
});
});
afterAll(() => {
Reflect.deleteProperty(globalThis, 'IS_REACT_ACT_ENVIRONMENT');
});
function Boom(): never {
throw new Error('render blew up');
}
/** Recursively clones the real exported route table, replacing only the
* `/chat` route's `element` with `<Boom />` — every other route (including
* the real `AuthGuard` nesting and the real `/chat` `errorElement`) is left
* exactly as exported. This is what makes the test fail if a future change
* removes the real route's `errorElement`, unlike a hand-built independent
* route tree that could drift from production undetected. */
function replaceChatElementWithBoom(nodes: RouteObject[]): RouteObject[] {
return nodes.map((node) => {
const cloned: RouteObject = { ...node };
if (cloned.path === '/chat') {
cloned.element = <Boom />;
}
if (cloned.children) {
cloned.children = replaceChatElementWithBoom(cloned.children);
}
return cloned;
});
}
let root: Root | null;
let container: HTMLElement;
afterEach(async () => {
await act(async () => {
root?.unmount();
});
document.body.replaceChildren();
root = null;
useSessionMock.mockReset();
});
describe('ChatRouteErrorBoundary', () => {
it('renders a recoverable, non-blank fallback when the /chat route element throws during render', async () => {
useSessionMock.mockReturnValue({ data: { user: { id: 'user-1' } }, isPending: false });
const routeObjects = replaceChatElementWithBoom(routes);
const router = createMemoryRouter(routeObjects, { initialEntries: ['/chat'] });
container = document.createElement('div');
document.body.append(container);
root = createRoot(container);
const consoleErrorSpy = vi.spyOn(console, 'error').mockImplementation(() => {});
try {
await act(async () => {
root?.render(<RouterProvider router={router} />);
});
expect(consoleErrorSpy).toHaveBeenCalled();
} finally {
consoleErrorSpy.mockRestore();
}
expect(container.textContent).not.toBe('');
expect(container.querySelector('[role="alert"]')).toBeTruthy();
});
});
@@ -0,0 +1,22 @@
import type { ReactElement } from 'react';
import { useRouteError } from 'react-router-dom';
/**
* `/chat` renders live, server-driven state (streamed text, tool calls,
* manifests) that can carry malformed payloads no compile-time contract can
* fully rule out at every dereference site. This is the last line of
* defense: if something still throws during render, show a recoverable
* alert instead of leaving the user on a blank/white screen.
*/
export function ChatRouteErrorBoundary(): ReactElement {
useRouteError();
return (
<div role="alert" className="flex min-h-screen flex-col items-center justify-center gap-3 p-8">
<p className="text-sm font-medium">Something went wrong loading chat.</p>
<a href="/chat" className="text-sm underline">
Reload chat
</a>
</div>
);
}
+636
View File
@@ -0,0 +1,636 @@
import { act } from 'react';
import { createRoot, type Root } from 'react-dom/client';
import { afterAll, afterEach, beforeAll, beforeEach, describe, expect, it, vi } from 'vitest';
import { createFakeChatSocket } from '@/spa/chat/test-support/fake-chat-socket';
import { MAX_MANIFEST_ITEMS } from '@/spa/chat/limits';
const { getSocketMock, destroySocketMock } = vi.hoisted(() => ({
getSocketMock: vi.fn(),
destroySocketMock: vi.fn(),
}));
vi.mock('@/lib/socket', () => ({
getSocket: getSocketMock,
destroySocket: destroySocketMock,
}));
import { ChatPage } from './chat';
function setValue(el: HTMLInputElement | HTMLTextAreaElement, value: string): void {
const proto =
el instanceof HTMLTextAreaElement ? HTMLTextAreaElement.prototype : HTMLInputElement.prototype;
const setter = Object.getOwnPropertyDescriptor(proto, 'value')?.set;
setter?.call(el, value);
el.dispatchEvent(new Event('input', { bubbles: true }));
}
function selectValue(el: HTMLSelectElement, value: string): void {
const setter = Object.getOwnPropertyDescriptor(HTMLSelectElement.prototype, 'value')?.set;
setter?.call(el, value);
el.dispatchEvent(new Event('change', { bubbles: true }));
}
function findButton(container: HTMLElement, text: string): HTMLButtonElement {
const button = [...container.querySelectorAll('button')].find((candidate) =>
candidate.textContent?.includes(text),
);
if (!button) throw new Error(`Button with text "${text}" not found`);
return button;
}
let fake: ReturnType<typeof createFakeChatSocket>;
let root: Root | null;
let container: HTMLElement;
beforeAll(() => {
Object.defineProperty(globalThis, 'IS_REACT_ACT_ENVIRONMENT', {
configurable: true,
value: true,
});
});
afterAll(() => {
Reflect.deleteProperty(globalThis, 'IS_REACT_ACT_ENVIRONMENT');
});
beforeEach(async () => {
fake = createFakeChatSocket();
getSocketMock.mockReset().mockReturnValue(fake.socket);
destroySocketMock.mockReset();
container = document.createElement('div');
document.body.append(container);
root = createRoot(container);
await act(async () => {
root?.render(<ChatPage />);
});
});
afterEach(async () => {
await act(async () => {
root?.unmount();
});
document.body.replaceChildren();
});
describe('ChatPage', () => {
it('streams agent:text and agent:thinking, shows tool status, and finalizes on agent:end with usage', async () => {
await act(async () => {
fake.serverEmit('message:ack', { conversationId: 'c1', messageId: 'm1' });
fake.serverEmit('agent:start', { conversationId: 'c1' });
});
await act(async () => {
fake.serverEmit('agent:thinking', { conversationId: 'c1', text: 'pondering…' });
fake.serverEmit('agent:text', { conversationId: 'c1', text: 'Hel' });
fake.serverEmit('agent:text', { conversationId: 'c1', text: 'lo!' });
fake.serverEmit('agent:tool:start', {
conversationId: 'c1',
toolCallId: 't1',
toolName: 'web_search',
});
});
expect(container.textContent).toContain('pondering…');
expect(container.textContent).toContain('Hello!');
expect(container.textContent).toContain('web_search');
expect(container.textContent).toMatch(/running/i);
await act(async () => {
fake.serverEmit('agent:tool:end', {
conversationId: 'c1',
toolCallId: 't1',
toolName: 'web_search',
isError: false,
});
fake.serverEmit('agent:end', {
conversationId: 'c1',
usage: {
provider: 'anthropic',
modelId: 'claude',
thinkingLevel: 'medium',
tokens: { input: 12, output: 34, cacheRead: 0, cacheWrite: 0, total: 46 },
cost: 0.02,
context: { percent: 3, window: 200000 },
},
});
});
expect(container.textContent).toMatch(/success/i);
expect(container.textContent).toContain('Hello!');
expect(container.textContent).toMatch(/46/);
});
it('renders the commands manifest and session info, and lets the user pick a thinking level', async () => {
await act(async () => {
fake.serverEmit('message:ack', { conversationId: 'c1', messageId: 'm1' });
fake.serverEmit('commands:manifest', {
manifest: {
commands: [
{
name: 'model',
aliases: ['m'],
description: 'Change the active model',
scope: 'core',
execution: 'socket',
available: true,
},
],
skills: [],
version: 1,
},
});
fake.serverEmit('session:info', {
conversationId: 'c1',
provider: 'anthropic',
modelId: 'claude',
thinkingLevel: 'medium',
availableThinkingLevels: ['low', 'medium', 'high'],
routingDecision: {
model: 'claude',
provider: 'anthropic',
ruleName: 'default',
reason: 'default routing',
},
});
});
expect(container.textContent).toContain('model');
expect(container.textContent).toContain('Change the active model');
expect(container.textContent).toContain('anthropic');
expect(container.textContent).toContain('default routing');
const select = container.querySelector(
'select[aria-label="Thinking level"]',
) as HTMLSelectElement;
expect(select).toBeTruthy();
expect([...select.options].map((o) => o.value)).toEqual(['low', 'medium', 'high']);
await act(async () => {
selectValue(select, 'high');
});
expect(fake.emitted).toContainEqual({
event: 'set:thinking',
payload: { conversationId: 'c1', level: 'high' },
});
});
it('executes and approves commands with exact payloads and surfaces the approval affordance', async () => {
await act(async () => {
fake.serverEmit('message:ack', { conversationId: 'c1', messageId: 'm1' });
});
const commandInput = container.querySelector(
'input[aria-label="Command name"]',
) as HTMLInputElement;
const argsInput = container.querySelector(
'input[aria-label="Command arguments"]',
) as HTMLInputElement;
await act(async () => {
setValue(commandInput, 'model');
setValue(argsInput, 'gpt-5');
});
await act(async () => {
findButton(container, 'Run command').click();
});
expect(fake.emitted).toContainEqual({
event: 'command:execute',
payload: { conversationId: 'c1', command: 'model', args: 'gpt-5' },
});
await act(async () => {
setValue(commandInput, 'deploy');
setValue(argsInput, 'prod');
});
await act(async () => {
findButton(container, 'Request approval').click();
});
expect(fake.emitted).toContainEqual({
event: 'command:approve',
payload: { conversationId: 'c1', command: 'deploy', args: 'prod' },
});
await act(async () => {
fake.serverEmit('command:approval', {
conversationId: 'c1',
command: 'deploy',
success: true,
approvalId: 'ap1',
expiresAt: '2026-01-01T00:00:00.000Z',
});
});
expect(container.textContent).toMatch(/approved/i);
await act(async () => {
findButton(container, 'Run approved command').click();
});
expect(fake.emitted).toContainEqual({
event: 'command:execute',
payload: { conversationId: 'c1', command: 'deploy', args: 'prod', approvalId: 'ap1' },
});
});
it('shows visible alert surfaces for a server error and the structured contract reason for a failed command result', async () => {
await act(async () => {
fake.serverEmit('message:ack', { conversationId: 'c1', messageId: 'm1' });
fake.serverEmit('error', { conversationId: 'c1', error: 'The model is unavailable' });
fake.serverEmit('command:result', {
conversationId: 'c1',
command: 'model',
success: false,
message: 'Unknown model',
});
});
const alerts = [...container.querySelectorAll('[role="alert"]')];
const alertText = alerts.map((node) => node.textContent).join(' ');
expect(alertText).toContain('The model is unavailable');
// The structured, contract-provided denial reason is visibly rendered.
expect(alertText).toContain('Unknown model');
});
it('falls back to a stable "Command failed." copy when a failed command result has no usable message', async () => {
await act(async () => {
fake.serverEmit('message:ack', { conversationId: 'c1', messageId: 'm1' });
fake.serverEmitRaw('command:result', {
conversationId: 'c1',
command: 'model',
success: false,
message: { bad: 'object' },
});
});
const alerts = [...container.querySelectorAll('[role="alert"]')];
const alertText = alerts.map((node) => node.textContent).join(' ');
expect(alertText).toContain('Command failed.');
});
it('caps availableThinkingLevels before storing and rendering a hostile session payload', async () => {
const hostileLevels = Array.from({ length: MAX_MANIFEST_ITEMS + 50 }, (_, i) => `level-${i}`);
await act(async () => {
fake.serverEmit('message:ack', { conversationId: 'c1', messageId: 'm1' });
fake.serverEmit('session:info', {
conversationId: 'c1',
provider: 'anthropic',
modelId: 'claude',
thinkingLevel: 'level-0',
availableThinkingLevels: hostileLevels,
});
});
const select = container.querySelector(
'select[aria-label="Thinking level"]',
) as HTMLSelectElement;
expect(select).toBeTruthy();
expect(select.options.length).toBeLessThanOrEqual(MAX_MANIFEST_ITEMS);
});
it('renders a safe fallback when session:info arrives with a malformed (non-array) availableThinkingLevels, without throwing', async () => {
await act(async () => {
fake.serverEmit('message:ack', { conversationId: 'c1', messageId: 'm1' });
fake.serverEmitRaw('session:info', {
conversationId: 'c1',
provider: 'anthropic',
modelId: 'claude',
thinkingLevel: 'medium',
availableThinkingLevels: null,
});
});
expect(container.querySelector('section[aria-label="Session info"]')).toBeTruthy();
const select = container.querySelector(
'select[aria-label="Thinking level"]',
) as HTMLSelectElement;
expect(select).toBeTruthy();
// A malformed level list still shows a visible, safe placeholder option
// rather than a silently empty select.
expect([...select.options]).toHaveLength(1);
expect(select.options[0]?.textContent).toMatch(/unavailable/i);
await act(async () => {
selectValue(select, '');
});
expect(fake.emitted.filter((e) => e.event === 'set:thinking')).toHaveLength(0);
});
it('renders honest unavailable labels — not fabricated zeros — when agent:end usage has malformed/missing numeric fields', async () => {
await act(async () => {
fake.serverEmit('message:ack', { conversationId: 'c1', messageId: 'm1' });
fake.serverEmit('agent:start', { conversationId: 'c1' });
fake.serverEmitRaw('agent:end', {
conversationId: 'c1',
usage: {
provider: { nested: 'object' },
modelId: undefined,
thinkingLevel: 'medium',
tokens: { total: 'not-a-number' },
cost: undefined,
context: { percent: null, window: 200000 },
},
});
});
const usage = container.querySelector('[aria-label="Usage"]');
expect(usage).toBeTruthy();
expect(usage?.textContent).toContain('tokens unavailable');
expect(usage?.textContent).toContain('cost unavailable');
expect(usage?.textContent).not.toContain('0 tokens');
expect(usage?.textContent).not.toContain('$0.0000');
expect(usage?.textContent).toContain('unknown/unknown');
});
it('renders a safe fallback for message:ack when messageId is a malformed non-string value, without throwing', async () => {
await expect(
act(async () => {
fake.serverEmitRaw('message:ack', { conversationId: 'c1', messageId: { bad: 'object' } });
}),
).resolves.not.toThrow();
const status = [...container.querySelectorAll('[role="status"]')].find((node) =>
node.textContent?.includes('Message accepted'),
);
expect(status).toBeTruthy();
// A malformed messageId gets a stable, visible fallback — never blank,
// never the raw object.
expect(status?.textContent).toContain('unknown');
});
it('renders safely and does not throw when system:reload.message is a malformed non-string value', async () => {
await expect(
act(async () => {
fake.serverEmitRaw('system:reload', {
commands: [],
skills: [],
providers: [],
message: { bad: 'object' },
});
}),
).resolves.not.toThrow();
const status = container.querySelector('[role="status"]');
expect(status).toBeTruthy();
// A malformed reload message renders a stable, visible fallback rather
// than a silently empty status line.
expect(status?.textContent).toContain('Commands reloaded.');
});
it('renders safely and does not throw when a scoped error carries a malformed non-string error value', async () => {
await expect(
act(async () => {
fake.serverEmit('message:ack', { conversationId: 'c1', messageId: 'm1' });
fake.serverEmitRaw('error', { conversationId: 'c1', error: ['not', 'a', 'string'] });
}),
).resolves.not.toThrow();
expect(container.querySelector('[role="alert"]')).toBeTruthy();
});
it('sends a message with optional provider/model fields and emits abort from the Stop control', async () => {
const textarea = container.querySelector(
'textarea[aria-label="Message"]',
) as HTMLTextAreaElement;
const providerInput = container.querySelector(
'input[aria-label="Provider"]',
) as HTMLInputElement;
const modelInput = container.querySelector('input[aria-label="Model"]') as HTMLInputElement;
const stopButtonBefore = container.querySelector(
'button[aria-label="Stop"]',
) as HTMLButtonElement;
expect(stopButtonBefore.disabled).toBe(true);
await act(async () => {
setValue(textarea, 'hello there');
setValue(providerInput, 'anthropic');
setValue(modelInput, 'claude');
});
await act(async () => {
textarea.dispatchEvent(
new KeyboardEvent('keydown', { key: 'Enter', bubbles: true, cancelable: true }),
);
});
expect(fake.emitted).toContainEqual({
event: 'message',
payload: {
conversationId: undefined,
content: 'hello there',
provider: 'anthropic',
modelId: 'claude',
},
});
expect(container.textContent).toContain('hello there');
await act(async () => {
fake.serverEmit('message:ack', { conversationId: 'c1', messageId: 'm1' });
fake.serverEmit('agent:start', { conversationId: 'c1' });
});
const stopButtonDuring = container.querySelector(
'button[aria-label="Stop"]',
) as HTMLButtonElement;
expect(stopButtonDuring.disabled).toBe(false);
await act(async () => {
stopButtonDuring.click();
});
expect(fake.emitted).toContainEqual({ event: 'abort', payload: { conversationId: 'c1' } });
});
it('renders the session panel from a pre-ack session:info and keeps it visible after the later ack', async () => {
const textarea = container.querySelector(
'textarea[aria-label="Message"]',
) as HTMLTextAreaElement;
await act(async () => {
setValue(textarea, 'hello');
});
await act(async () => {
textarea.dispatchEvent(
new KeyboardEvent('keydown', { key: 'Enter', bubbles: true, cancelable: true }),
);
});
await act(async () => {
fake.serverEmit('session:info', {
conversationId: 'c1',
provider: 'anthropic',
modelId: 'claude',
thinkingLevel: 'medium',
availableThinkingLevels: ['low', 'medium', 'high'],
});
});
expect(container.querySelector('section[aria-label="Session info"]')).toBeTruthy();
expect(container.textContent).toContain('anthropic');
await act(async () => {
fake.serverEmit('message:ack', { conversationId: 'c1', messageId: 'm1' });
});
expect(container.querySelector('section[aria-label="Session info"]')).toBeTruthy();
expect(container.textContent).toContain('anthropic');
});
it('surfaces a pre-ack error as an alert without leaving the Stop control stuck active', async () => {
const textarea = container.querySelector(
'textarea[aria-label="Message"]',
) as HTMLTextAreaElement;
await act(async () => {
setValue(textarea, 'hello');
});
await act(async () => {
textarea.dispatchEvent(
new KeyboardEvent('keydown', { key: 'Enter', bubbles: true, cancelable: true }),
);
});
await act(async () => {
fake.serverEmit('error', {
conversationId: 'c1',
error: 'Failed to start agent session. Please try again.',
});
});
const alerts = [...container.querySelectorAll('[role="alert"]')];
expect(alerts.some((node) => node.textContent?.includes('Failed to start agent session'))).toBe(
true,
);
const stopButton = container.querySelector('button[aria-label="Stop"]') as HTMLButtonElement;
expect(stopButton.disabled).toBe(true);
});
it('shows an accessible status once the message is acknowledged', async () => {
await act(async () => {
fake.serverEmit('message:ack', { conversationId: 'c1', messageId: 'm1' });
});
const statuses = [...container.querySelectorAll('[role="status"]')];
expect(statuses.some((node) => node.textContent?.includes('m1'))).toBe(true);
});
it('renders finalized thinking text in the transcript after agent:end, not only while streaming', async () => {
await act(async () => {
fake.serverEmit('message:ack', { conversationId: 'c1', messageId: 'm1' });
fake.serverEmit('agent:start', { conversationId: 'c1' });
fake.serverEmit('agent:thinking', { conversationId: 'c1', text: 'reasoning about it' });
fake.serverEmit('agent:text', { conversationId: 'c1', text: 'Done.' });
});
expect(container.textContent).toContain('reasoning about it');
await act(async () => {
fake.serverEmit('agent:end', { conversationId: 'c1' });
});
expect(container.textContent).toContain('reasoning about it');
expect(container.textContent).toContain('Done.');
});
it('ignores a concurrent approval request and only executes the approved command once', async () => {
await act(async () => {
fake.serverEmit('message:ack', { conversationId: 'c1', messageId: 'm1' });
});
const commandInput = container.querySelector(
'input[aria-label="Command name"]',
) as HTMLInputElement;
const argsInput = container.querySelector(
'input[aria-label="Command arguments"]',
) as HTMLInputElement;
await act(async () => {
setValue(commandInput, 'deploy');
setValue(argsInput, 'prod');
});
await act(async () => {
findButton(container, 'Request approval').click();
});
await act(async () => {
setValue(argsInput, 'staging');
});
await act(async () => {
findButton(container, 'Request approval').click();
});
expect(fake.emitted.filter((e) => e.event === 'command:approve')).toHaveLength(1);
expect(fake.emitted).toContainEqual({
event: 'command:approve',
payload: { conversationId: 'c1', command: 'deploy', args: 'prod' },
});
await act(async () => {
fake.serverEmit('command:approval', {
conversationId: 'c1',
command: 'deploy',
success: true,
approvalId: 'ap1',
expiresAt: '2026-01-01T00:00:00.000Z',
});
});
await act(async () => {
findButton(container, 'Run approved command').click();
findButton(container, 'Run approved command').click();
});
expect(fake.emitted.filter((e) => e.event === 'command:execute')).toHaveLength(1);
expect(fake.emitted).toContainEqual({
event: 'command:execute',
payload: { conversationId: 'c1', command: 'deploy', args: 'prod', approvalId: 'ap1' },
});
});
it('disables sending a second message while a turn is streaming', async () => {
const textarea = container.querySelector(
'textarea[aria-label="Message"]',
) as HTMLTextAreaElement;
await act(async () => {
setValue(textarea, 'first');
});
await act(async () => {
textarea.dispatchEvent(
new KeyboardEvent('keydown', { key: 'Enter', bubbles: true, cancelable: true }),
);
});
await act(async () => {
fake.serverEmit('message:ack', { conversationId: 'c1', messageId: 'm1' });
fake.serverEmit('agent:start', { conversationId: 'c1' });
});
const sendButton = findButton(container, 'Send');
expect(sendButton.disabled).toBe(true);
await act(async () => {
setValue(textarea, 'second');
});
await act(async () => {
textarea.dispatchEvent(
new KeyboardEvent('keydown', { key: 'Enter', bubbles: true, cancelable: true }),
);
});
expect(fake.emitted.filter((e) => e.event === 'message')).toHaveLength(1);
});
it('removes socket handlers and tears down the socket on unmount, with no network calls', async () => {
expect(fake.listeners.size).toBeGreaterThan(0);
await act(async () => {
root?.unmount();
});
root = null;
for (const [, handlers] of fake.listeners) {
expect(handlers.size).toBe(0);
}
expect(destroySocketMock).toHaveBeenCalledOnce();
});
});
+92
View File
@@ -0,0 +1,92 @@
import type { ReactElement } from 'react';
import { CommandsPanel } from '@/spa/chat/commands-panel';
import { Composer } from '@/spa/chat/composer';
import { MessageTranscript } from '@/spa/chat/message-transcript';
import { asFiniteNumberOrNull, asString } from '@/spa/chat/runtime-guards';
import { SessionPanel } from '@/spa/chat/session-panel';
import { ToolCallList } from '@/spa/chat/tool-call-list';
import { useChatConnection } from '@/spa/chat/use-chat-connection';
/** Renders a real value normally, but an honest "unavailable" label instead
* of a fabricated `0` for a missing/malformed count a real `0 tokens` and
* an unknown token count must never look the same. */
function formatTokens(value: unknown): string {
const tokens = asFiniteNumberOrNull(value);
return tokens === null ? 'tokens unavailable' : `${tokens} tokens`;
}
/** Same honesty guarantee as `formatTokens`, for cost. */
function formatCost(value: unknown): string {
const cost = asFiniteNumberOrNull(value);
return cost === null ? 'cost unavailable' : `$${cost.toFixed(4)}`;
}
export function ChatPage(): ReactElement {
const { state, actions } = useChatConnection();
const hasConversation = state.conversationId !== null;
return (
<div className="flex h-[calc(100vh-3.5rem)] min-h-0 flex-col overflow-hidden md:h-screen">
<header className="border-b px-4 py-3">
<h1 className="text-lg font-semibold">Chat</h1>
</header>
{state.systemReload ? (
<div role="status" className="border-b px-4 py-2 text-sm">
{asString(state.systemReload.message)}
</div>
) : null}
{state.error ? (
<div role="alert" className="border-b px-4 py-2 text-sm">
{asString(state.error)}
</div>
) : null}
{state.ack ? (
<div role="status" className="border-b px-4 py-1 text-xs opacity-70">
Message accepted · conversation {asString(state.ack.conversationId, 'unknown')} · id{' '}
{asString(state.ack.messageId, 'unknown')}
</div>
) : null}
<SessionPanel sessionInfo={state.sessionInfo} onSetThinking={actions.setThinking} />
<MessageTranscript messages={state.messages} streaming={state.streaming} text={state.text} />
{state.thinking ? (
<section aria-label="Thinking" className="px-4 pb-2 text-xs italic opacity-80">
{state.thinking}
</section>
) : null}
<ToolCallList tools={state.tools} />
{state.usage ? (
<div aria-label="Usage" className="px-4 pb-2 text-xs opacity-80">
{formatTokens(state.usage.tokens?.total)} · {formatCost(state.usage.cost)} ·{' '}
{asString(state.usage.provider, 'unknown')}/{asString(state.usage.modelId, 'unknown')}
</div>
) : null}
<CommandsPanel
manifest={state.manifest}
results={state.commandResults}
approval={state.approval}
pendingApproval={state.pendingApproval}
hasConversation={hasConversation}
onExecute={actions.executeCommand}
onApprove={actions.approveCommand}
onRunApproved={actions.runApprovedCommand}
/>
<Composer
onSend={actions.sendMessage}
onStop={actions.abort}
streaming={state.streaming}
sending={state.sending}
hasConversation={hasConversation}
/>
</div>
);
}
+7
View File
@@ -0,0 +1,7 @@
export function getErrorMessage(error: unknown, fallback: string): string {
if (error instanceof Error && error.message.trim().length > 0) {
return error.message;
}
return fallback;
}
+115
View File
@@ -0,0 +1,115 @@
import type { Mission, Project, Task } from '@/lib/types';
export const projectFixtures: Project[] = [
{
id: 'project-1',
name: 'Mosaic Stack',
description: 'Gateway and dashboard parity work',
status: 'active',
userId: 'user-1',
metadata: {
prd: '# Mosaic Stack PRD\n\n## Objective\n\nShip the SPA route parity pages.',
},
createdAt: '2026-08-01T12:00:00.000Z',
updatedAt: '2026-08-09T18:30:00.000Z',
},
{
id: 'project-2',
name: 'Agent Runtime',
description: 'Pi SDK integration',
status: 'paused',
userId: 'user-1',
metadata: null,
createdAt: '2026-08-02T08:00:00.000Z',
updatedAt: '2026-08-05T10:00:00.000Z',
},
];
export const missionFixtures: Mission[] = [
{
id: 'mission-1',
name: 'Ship web parity',
description: 'Port read-only routes into the SPA',
status: 'active',
projectId: 'project-1',
metadata: null,
createdAt: '2026-08-03T12:00:00.000Z',
updatedAt: '2026-08-09T12:00:00.000Z',
},
{
id: 'mission-2',
name: 'Unrelated mission',
description: 'Must be filtered out of the project detail view',
status: 'planning',
projectId: 'project-2',
metadata: null,
createdAt: '2026-08-04T12:00:00.000Z',
updatedAt: '2026-08-04T12:00:00.000Z',
},
];
export const taskFixtures: Task[] = [
{
id: 'task-1',
title: 'Route /projects',
description: 'Port the read-only projects listing into the SPA',
status: 'done',
priority: 'high',
projectId: 'project-1',
missionId: 'mission-1',
assignee: 'Jarvis',
tags: ['spa', 'projects'],
dueDate: '2026-08-12T00:00:00.000Z',
metadata: {
notes: 'Read-only modal content should remain intact.',
pr_links: [{ url: 'https://example.invalid/pr/1', label: 'PR #1' }],
},
createdAt: '2026-08-04T10:00:00.000Z',
updatedAt: '2026-08-09T15:00:00.000Z',
},
{
id: 'task-2',
title: 'Route /projects/:id',
description: 'Reuse overview, tasks, missions, and PRD tabs',
status: 'in-progress',
priority: 'critical',
projectId: 'project-1',
missionId: 'mission-1',
assignee: null,
tags: ['spa', 'detail'],
dueDate: null,
metadata: null,
createdAt: '2026-08-05T09:00:00.000Z',
updatedAt: '2026-08-10T08:00:00.000Z',
},
{
id: 'task-3',
title: 'Route /tasks',
description: 'Wire list and kanban modal interactions',
status: 'blocked',
priority: 'medium',
projectId: 'project-1',
missionId: null,
assignee: null,
tags: ['spa', 'tasks'],
dueDate: null,
metadata: null,
createdAt: '2026-08-06T09:00:00.000Z',
updatedAt: '2026-08-08T08:00:00.000Z',
},
{
id: 'task-4',
title: 'Other project task',
description: 'Used only to confirm mission filtering remains project-scoped',
status: 'not-started',
priority: 'low',
projectId: 'project-2',
missionId: 'mission-2',
assignee: null,
tags: null,
dueDate: null,
metadata: null,
createdAt: '2026-08-06T09:00:00.000Z',
updatedAt: '2026-08-06T09:00:00.000Z',
},
];
@@ -0,0 +1,174 @@
import { act } from 'react';
import { createRoot, type Root } from 'react-dom/client';
import { createMemoryRouter, RouterProvider, type RouteObject } from 'react-router-dom';
import { afterAll, afterEach, beforeAll, describe, expect, it, vi } from 'vitest';
import { missionFixtures, projectFixtures, taskFixtures } from './page-fixtures';
const { apiMock } = vi.hoisted(() => ({
apiMock: vi.fn(),
}));
vi.mock('@/lib/api', () => ({
api: apiMock,
}));
import { ProjectDetailPage } from './project-detail';
let root: Root | null = null;
let container: HTMLDivElement;
beforeAll(() => {
Object.defineProperty(globalThis, 'IS_REACT_ACT_ENVIRONMENT', {
configurable: true,
value: true,
});
});
afterAll(() => {
Reflect.deleteProperty(globalThis, 'IS_REACT_ACT_ENVIRONMENT');
});
afterEach(async () => {
await act(async () => {
root?.unmount();
});
document.body.replaceChildren();
root = null;
apiMock.mockReset();
});
async function renderProjectDetailPage(): Promise<ReturnType<typeof createMemoryRouter>> {
const routes: RouteObject[] = [
{ path: '/projects', element: <p>Projects index target</p> },
{ path: '/projects/:id', element: <ProjectDetailPage /> },
];
const router = createMemoryRouter(routes, { initialEntries: ['/projects/project-1'] });
container = document.createElement('div');
document.body.append(container);
root = createRoot(container);
await act(async () => {
root?.render(<RouterProvider router={router} />);
});
return router;
}
function clickButtonByText(text: string): void {
const button = [...container.querySelectorAll('button')].find((candidate) =>
candidate.textContent?.includes(text),
);
if (!button) {
throw new Error(`Button containing "${text}" not found`);
}
button.dispatchEvent(new MouseEvent('click', { bubbles: true }));
}
describe('ProjectDetailPage', () => {
it('loads the project, tasks, missions, and optional PRD content for the active project', async () => {
apiMock
.mockResolvedValueOnce(projectFixtures[0])
.mockResolvedValueOnce(missionFixtures)
.mockResolvedValueOnce(taskFixtures.filter((task) => task.projectId === 'project-1'));
await renderProjectDetailPage();
expect(apiMock.mock.calls).toEqual([
['/api/projects/project-1'],
['/api/missions'],
['/api/tasks?projectId=project-1'],
]);
expect(container.textContent).toContain('Mosaic Stack');
expect(container.textContent).toContain('Route /projects/:id');
expect(container.textContent).toContain('Tasks');
expect(container.textContent).toContain('Done');
expect(container.textContent).toContain('Blocked');
await act(async () => {
clickButtonByText('Missions (1)');
});
expect(container.textContent).toContain('Ship web parity');
expect(container.textContent).not.toContain('Unrelated mission');
await act(async () => {
clickButtonByText('PRD');
});
expect(container.textContent).toContain('Mosaic Stack PRD');
expect(container.textContent).toContain('Ship the SPA route parity pages.');
});
it('opens and closes the existing read-only task modal from the tasks tab', async () => {
apiMock
.mockResolvedValueOnce(projectFixtures[0])
.mockResolvedValueOnce(missionFixtures)
.mockResolvedValueOnce(taskFixtures.filter((task) => task.projectId === 'project-1'));
await renderProjectDetailPage();
await act(async () => {
clickButtonByText('Tasks (3)');
});
const row = [...container.querySelectorAll('tr')].find((candidate) =>
candidate.textContent?.includes('Route /projects'),
);
expect(row).toBeTruthy();
await act(async () => {
row?.dispatchEvent(new MouseEvent('click', { bubbles: true }));
});
expect(container.querySelector('[role="dialog"]')).toBeTruthy();
expect(container.textContent).toContain('Read-only modal content should remain intact.');
const closeButton = container.querySelector('button[aria-label="Close task details"]');
expect(closeButton).toBeTruthy();
await act(async () => {
closeButton?.dispatchEvent(new MouseEvent('click', { bubbles: true }));
});
expect(container.querySelector('[role="dialog"]')).toBeNull();
});
it('renders the project with an empty missions tab when the missions request fails', async () => {
apiMock
.mockResolvedValueOnce(projectFixtures[0])
.mockRejectedValueOnce(new Error('Missions request failed'))
.mockResolvedValueOnce(taskFixtures.filter((task) => task.projectId === 'project-1'));
await renderProjectDetailPage();
expect(container.textContent).toContain('Mosaic Stack');
expect(container.querySelector('[role="alert"]')).toBeNull();
await act(async () => {
clickButtonByText('Missions (0)');
});
expect(container.textContent).toContain('No missions for this project');
});
it('renders a visible alert when the project request fails and lets the user navigate back', async () => {
apiMock
.mockRejectedValueOnce(new Error('Project request failed'))
.mockResolvedValueOnce(missionFixtures)
.mockResolvedValueOnce(taskFixtures.filter((task) => task.projectId === 'project-1'));
const router = await renderProjectDetailPage();
const alert = container.querySelector('[role="alert"]');
expect(alert).toBeTruthy();
expect(alert?.textContent).toContain('Project request failed');
expect(container.textContent).not.toContain('Mosaic Stack');
await act(async () => {
clickButtonByText('Back to projects');
});
expect(router.state.location.pathname).toBe('/projects');
});
});
+344
View File
@@ -0,0 +1,344 @@
import { useEffect, useState, type ReactElement } from 'react';
import { useNavigate, useParams } from 'react-router-dom';
import { MissionTimeline } from '@/components/projects/mission-timeline';
import { PrdViewer } from '@/components/projects/prd-viewer';
import { TaskDetailModal } from '@/components/tasks/task-detail-modal';
import { TaskListView } from '@/components/tasks/task-list-view';
import { TaskStatusSummary } from '@/components/tasks/task-status-summary';
import { api } from '@/lib/api';
import { cn } from '@/lib/cn';
import type { Mission, Project, Task, TaskStatus } from '@/lib/types';
import { getErrorMessage } from './page-errors';
type Tab = 'overview' | 'tasks' | 'missions' | 'prd';
const projectStatusColors: Record<string, string> = {
active: 'bg-success/20 text-success',
paused: 'bg-warning/20 text-warning',
completed: 'bg-blue-600/20 text-blue-400',
archived: 'bg-gray-600/20 text-gray-400',
};
const taskStatusColors: Record<string, string> = {
'not-started': 'bg-gray-600/20 text-gray-300',
'in-progress': 'bg-blue-600/20 text-blue-400',
blocked: 'bg-error/20 text-error',
done: 'bg-success/20 text-success',
cancelled: 'bg-gray-600/20 text-gray-500',
};
interface TabButtonProps {
id: Tab;
label: string;
activeTab: Tab;
onClick: (tab: Tab) => void;
}
function TabButton({ id, label, activeTab, onClick }: TabButtonProps): ReactElement {
return (
<button
type="button"
onClick={() => onClick(id)}
className={cn(
'border-b-2 px-4 py-2 text-sm transition-colors',
activeTab === id
? 'border-text-primary text-text-primary'
: 'border-transparent text-text-muted hover:text-text-secondary',
)}
>
{label}
</button>
);
}
export function ProjectDetailPage(): ReactElement {
const { id = '' } = useParams();
const navigate = useNavigate();
const [project, setProject] = useState<Project | null>(null);
const [missions, setMissions] = useState<Mission[]>([]);
const [tasks, setTasks] = useState<Task[]>([]);
const [loading, setLoading] = useState(true);
const [error, setError] = useState<string | null>(null);
const [activeTab, setActiveTab] = useState<Tab>('overview');
const [taskFilter, setTaskFilter] = useState<TaskStatus | 'all'>('all');
const [selectedTask, setSelectedTask] = useState<Task | null>(null);
useEffect(() => {
if (!id) {
setError('Project id is missing.');
setLoading(false);
return;
}
let cancelled = false;
setLoading(true);
setError(null);
void Promise.all([
api<Project>('/api/projects/' + id),
api<Mission[]>('/api/missions').catch(() => [] as Mission[]),
api<Task[]>('/api/tasks?projectId=' + id).catch(() => [] as Task[]),
])
.then(([loadedProject, allMissions, loadedTasks]) => {
if (cancelled) return;
setProject(loadedProject);
setMissions(allMissions.filter((mission) => mission.projectId === id));
setTasks(loadedTasks);
})
.catch((caught: unknown) => {
if (cancelled) return;
setError(getErrorMessage(caught, 'Failed to load project.'));
})
.finally(() => {
if (cancelled) return;
setLoading(false);
});
return () => {
cancelled = true;
};
}, [id]);
if (loading) {
return (
<div className="flex min-h-screen flex-col px-4 py-6 sm:px-6">
<header className="mb-6 border-b px-1 pb-3">
<h1 className="text-2xl font-semibold">Project</h1>
</header>
<p className="py-16 text-center text-sm text-text-muted">Loading project...</p>
</div>
);
}
if (error || !project) {
return (
<div className="flex min-h-screen flex-col px-4 py-6 sm:px-6">
<header className="mb-6 border-b px-1 pb-3">
<h1 className="text-2xl font-semibold">Project</h1>
</header>
<div role="alert" className="rounded-lg border border-error/40 px-4 py-3 text-sm">
{error ?? 'Project not found.'}
</div>
<button
type="button"
onClick={() => navigate('/projects')}
className="mt-4 w-fit text-sm underline"
>
Back to projects
</button>
</div>
);
}
const filteredTasks =
taskFilter === 'all' ? tasks : tasks.filter((task) => task.status === taskFilter);
const prdContent = getPrdContent(project);
const tabs: Array<{ id: Tab; label: string }> = [
{ id: 'overview', label: 'Overview' },
{ id: 'tasks', label: `Tasks (${tasks.length})` },
{ id: 'missions', label: `Missions (${missions.length})` },
...(prdContent ? [{ id: 'prd' as const, label: 'PRD' }] : []),
];
return (
<div className="flex min-h-screen flex-col px-4 py-6 sm:px-6">
<header className="mb-6 border-b px-1 pb-3">
<nav className="mb-4 flex items-center gap-2 text-sm text-text-muted">
<button
type="button"
onClick={() => navigate('/projects')}
className="hover:text-text-secondary"
>
Projects
</button>
<span>/</span>
<span className="text-text-primary">{project.name}</span>
</nav>
<div className="flex items-start justify-between gap-4">
<div>
<div className="flex items-center gap-3">
<h1 className="text-2xl font-semibold text-text-primary">{project.name}</h1>
<span
className={cn(
'rounded-full px-2 py-0.5 text-xs',
projectStatusColors[project.status] ?? 'bg-gray-600/20 text-gray-400',
)}
>
{project.status}
</span>
</div>
{project.description ? (
<p className="mt-1 text-sm text-text-muted">{project.description}</p>
) : null}
<p className="mt-2 text-xs text-text-muted">
Created {new Date(project.createdAt).toLocaleDateString()} · Updated{' '}
{new Date(project.updatedAt).toLocaleDateString()}
</p>
</div>
</div>
</header>
<div className="mb-6 grid grid-cols-2 gap-3 sm:grid-cols-4">
<StatCard label="Tasks" value={String(tasks.length)} />
<StatCard
label="Done"
value={String(tasks.filter((task) => task.status === 'done').length)}
valueClass="text-success"
/>
<StatCard
label="In Progress"
value={String(tasks.filter((task) => task.status === 'in-progress').length)}
valueClass="text-blue-400"
/>
<StatCard
label="Blocked"
value={String(tasks.filter((task) => task.status === 'blocked').length)}
valueClass={tasks.some((task) => task.status === 'blocked') ? 'text-error' : undefined}
/>
</div>
<div className="mb-6 flex gap-0 border-b border-surface-border">
{tabs.map((tab) => (
<TabButton
key={tab.id}
id={tab.id}
label={tab.label}
activeTab={activeTab}
onClick={setActiveTab}
/>
))}
</div>
{activeTab === 'overview' ? (
<OverviewTab project={project} missions={missions} tasks={tasks} />
) : null}
{activeTab === 'tasks' ? (
<div>
<div className="mb-4">
<TaskStatusSummary
tasks={tasks}
activeFilter={taskFilter}
onFilterChange={setTaskFilter}
/>
</div>
<TaskListView tasks={filteredTasks} onTaskClick={setSelectedTask} />
</div>
) : null}
{activeTab === 'missions' ? <MissionTimeline missions={missions} /> : null}
{activeTab === 'prd' && prdContent ? (
<div className="rounded-lg border border-surface-border bg-surface-card p-6">
<PrdViewer content={prdContent} />
</div>
) : null}
{selectedTask ? (
<TaskDetailModal task={selectedTask} onClose={() => setSelectedTask(null)} />
) : null}
</div>
);
}
function OverviewTab({
project,
missions,
tasks,
}: {
project: Project;
missions: Mission[];
tasks: Task[];
}): ReactElement {
const recentTasks = [...tasks]
.sort((left, right) => new Date(right.updatedAt).getTime() - new Date(left.updatedAt).getTime())
.slice(0, 5);
return (
<div className="grid gap-6 lg:grid-cols-2">
<section>
<h2 className="mb-3 text-sm font-semibold text-text-secondary">Recent Tasks</h2>
{recentTasks.length === 0 ? (
<div className="rounded-lg border border-surface-border bg-surface-card p-4 text-center">
<p className="text-sm text-text-muted">No tasks yet</p>
</div>
) : (
<div className="space-y-2">
{recentTasks.map((task) => (
<div
key={task.id}
className="flex items-center justify-between gap-2 rounded-lg border border-surface-border bg-surface-card px-3 py-2"
>
<span className="truncate text-sm text-text-primary">{task.title}</span>
<span
className={cn(
'shrink-0 rounded-full px-2 py-0.5 text-xs',
taskStatusColors[task.status] ?? 'bg-gray-600/20 text-gray-400',
)}
>
{task.status}
</span>
</div>
))}
</div>
)}
</section>
<section>
<h2 className="mb-3 text-sm font-semibold text-text-secondary">Missions</h2>
{missions.length === 0 ? (
<div className="rounded-lg border border-surface-border bg-surface-card p-4 text-center">
<p className="text-sm text-text-muted">No missions yet</p>
</div>
) : (
<MissionTimeline missions={missions.slice(0, 4)} />
)}
</section>
{project.metadata && Object.keys(project.metadata).length > 0 ? (
<section className="lg:col-span-2">
<h2 className="mb-3 text-sm font-semibold text-text-secondary">Project Metadata</h2>
<div className="rounded-lg border border-surface-border bg-surface-card p-4">
<pre className="overflow-x-auto text-xs text-text-muted">
{JSON.stringify(project.metadata, null, 2)}
</pre>
</div>
</section>
) : null}
</div>
);
}
function StatCard({
label,
value,
valueClass,
}: {
label: string;
value: string;
valueClass?: string;
}): ReactElement {
return (
<div className="rounded-lg border border-surface-border bg-surface-card p-3">
<p className="text-xs text-text-muted">{label}</p>
<p className={cn('mt-1 text-lg font-semibold', valueClass ?? 'text-text-primary')}>{value}</p>
</div>
);
}
function getPrdContent(project: Project): string | null {
if (!project.metadata) return null;
const prd = project.metadata['prd'];
if (typeof prd === 'string' && prd.trim().length > 0) {
return prd;
}
const prdContent = project.metadata['prdContent'];
if (typeof prdContent === 'string' && prdContent.trim().length > 0) {
return prdContent;
}
return null;
}
+131
View File
@@ -0,0 +1,131 @@
import { act } from 'react';
import { createRoot, type Root } from 'react-dom/client';
import { createMemoryRouter, RouterProvider, type RouteObject } from 'react-router-dom';
import { afterAll, afterEach, beforeAll, describe, expect, it, vi } from 'vitest';
import { projectFixtures } from './page-fixtures';
const { apiMock } = vi.hoisted(() => ({
apiMock: vi.fn(),
}));
vi.mock('@/lib/api', () => ({
api: apiMock,
}));
import { ProjectsPage } from './projects';
interface Deferred<T> {
promise: Promise<T>;
resolve: (value: T) => void;
reject: (reason?: unknown) => void;
}
function createDeferred<T>(): Deferred<T> {
let resolve!: (value: T) => void;
let reject!: (reason?: unknown) => void;
const promise = new Promise<T>((res, rej) => {
resolve = res;
reject = rej;
});
return { promise, resolve, reject };
}
let root: Root | null = null;
let container: HTMLDivElement;
beforeAll(() => {
Object.defineProperty(globalThis, 'IS_REACT_ACT_ENVIRONMENT', {
configurable: true,
value: true,
});
});
afterAll(() => {
Reflect.deleteProperty(globalThis, 'IS_REACT_ACT_ENVIRONMENT');
});
afterEach(async () => {
await act(async () => {
root?.unmount();
});
document.body.replaceChildren();
root = null;
apiMock.mockReset();
});
async function renderProjectsPage(): Promise<ReturnType<typeof createMemoryRouter>> {
const routes: RouteObject[] = [
{ path: '/projects', element: <ProjectsPage /> },
{ path: '/projects/:id', element: <p>Project detail target</p> },
];
const router = createMemoryRouter(routes, { initialEntries: ['/projects'] });
container = document.createElement('div');
document.body.append(container);
root = createRoot(container);
await act(async () => {
root?.render(<RouterProvider router={router} />);
});
return router;
}
describe('ProjectsPage', () => {
it('shows a visible loading state while the project request is in flight', async () => {
const deferred = createDeferred<typeof projectFixtures>();
apiMock.mockReturnValueOnce(deferred.promise);
await renderProjectsPage();
expect(container.textContent).toContain('Loading projects...');
await act(async () => {
deferred.resolve(projectFixtures);
await deferred.promise;
});
});
it('renders project cards from the API and navigates to a project detail route on click', async () => {
apiMock.mockResolvedValueOnce(projectFixtures);
const router = await renderProjectsPage();
expect(apiMock).toHaveBeenCalledWith('/api/projects');
expect(container.textContent).toContain('Mosaic Stack');
expect(container.textContent).toContain('Agent Runtime');
const button = [...container.querySelectorAll('button')].find((candidate) =>
candidate.textContent?.includes('Mosaic Stack'),
);
expect(button).toBeTruthy();
await act(async () => {
button?.dispatchEvent(new MouseEvent('click', { bubbles: true }));
});
expect(router.state.location.pathname).toBe('/projects/project-1');
expect(container.textContent).toContain('Project detail target');
});
it('renders the empty state when the API returns no projects', async () => {
apiMock.mockResolvedValueOnce([]);
await renderProjectsPage();
expect(container.textContent).toContain('No projects yet');
expect(container.textContent).toContain(
'Projects will appear here when created via the gateway API',
);
});
it('renders a visible alert when the projects request fails', async () => {
apiMock.mockRejectedValueOnce(new Error('Projects are unavailable'));
await renderProjectsPage();
const alert = container.querySelector('[role="alert"]');
expect(alert).toBeTruthy();
expect(alert?.textContent).toContain('Projects are unavailable');
});
});
+70
View File
@@ -0,0 +1,70 @@
import { useEffect, useState, type ReactElement } from 'react';
import { useNavigate } from 'react-router-dom';
import { ProjectCard } from '@/components/projects/project-card';
import { api } from '@/lib/api';
import type { Project } from '@/lib/types';
import { getErrorMessage } from './page-errors';
export function ProjectsPage(): ReactElement {
const navigate = useNavigate();
const [projects, setProjects] = useState<Project[]>([]);
const [loading, setLoading] = useState(true);
const [error, setError] = useState<string | null>(null);
useEffect(() => {
let cancelled = false;
void api<Project[]>('/api/projects')
.then((response) => {
if (cancelled) return;
setProjects(response);
})
.catch((caught: unknown) => {
if (cancelled) return;
setError(getErrorMessage(caught, 'Failed to load projects.'));
})
.finally(() => {
if (cancelled) return;
setLoading(false);
});
return () => {
cancelled = true;
};
}, []);
return (
<div className="flex min-h-screen flex-col px-4 py-6 sm:px-6">
<header className="mb-6 border-b px-1 pb-3">
<h1 className="text-2xl font-semibold">Projects</h1>
</header>
{error ? (
<div role="alert" className="mb-6 rounded-lg border border-error/40 px-4 py-3 text-sm">
{error}
</div>
) : null}
{loading ? (
<p className="py-8 text-center text-sm text-text-muted">Loading projects...</p>
) : projects.length === 0 ? (
<div className="py-12 text-center">
<h2 className="text-lg font-medium text-text-secondary">No projects yet</h2>
<p className="mt-1 text-sm text-text-muted">
Projects will appear here when created via the gateway API
</p>
</div>
) : (
<div className="grid gap-4 sm:grid-cols-2 lg:grid-cols-3">
{projects.map((project) => (
<ProjectCard
key={project.id}
project={project}
onClick={(selectedProject) => navigate(`/projects/${selectedProject.id}`)}
/>
))}
</div>
)}
</div>
);
}
@@ -0,0 +1,86 @@
import { act } from 'react';
import { createRoot, type Root } from 'react-dom/client';
import { createMemoryRouter, RouterProvider, type RouteObject } from 'react-router-dom';
import { afterAll, afterEach, beforeAll, describe, expect, it, vi } from 'vitest';
const { useSessionMock } = vi.hoisted(() => ({
useSessionMock: vi.fn(),
}));
vi.mock('@/lib/auth-client', () => ({
useSession: useSessionMock,
}));
import { routes } from '@/routes';
function Boom(): never {
throw new Error('resource route render blew up');
}
function replaceRouteElementWithBoom(nodes: RouteObject[], path: string): RouteObject[] {
return nodes.map((node) => {
const cloned: RouteObject = { ...node };
if (cloned.path === path) {
cloned.element = <Boom />;
}
if (cloned.children) {
cloned.children = replaceRouteElementWithBoom(cloned.children, path);
}
return cloned;
});
}
let root: Root | null = null;
let container: HTMLDivElement;
beforeAll(() => {
Object.defineProperty(globalThis, 'IS_REACT_ACT_ENVIRONMENT', {
configurable: true,
value: true,
});
});
afterAll(() => {
Reflect.deleteProperty(globalThis, 'IS_REACT_ACT_ENVIRONMENT');
});
afterEach(async () => {
await act(async () => {
root?.unmount();
});
document.body.replaceChildren();
root = null;
useSessionMock.mockReset();
});
describe('resource route error boundaries', () => {
it.each(['/projects', '/projects/:id', '/tasks'])(
'renders a recoverable fallback when %s throws during route render',
async (path) => {
useSessionMock.mockReturnValue({ data: { user: { id: 'user-1' } }, isPending: false });
const initialEntry = path === '/projects/:id' ? '/projects/project-1' : path;
const router = createMemoryRouter(replaceRouteElementWithBoom(routes, path), {
initialEntries: [initialEntry],
});
container = document.createElement('div');
document.body.append(container);
root = createRoot(container);
const consoleErrorSpy = vi.spyOn(console, 'error').mockImplementation(() => {});
try {
await act(async () => {
root?.render(<RouterProvider router={router} />);
});
expect(consoleErrorSpy).toHaveBeenCalled();
} finally {
consoleErrorSpy.mockRestore();
}
expect(container.textContent).not.toBe('');
expect(container.querySelector('[role="alert"]')).toBeTruthy();
},
);
});
@@ -0,0 +1,55 @@
import type { ReactElement } from 'react';
import { useRouteError } from 'react-router-dom';
interface ResourceRouteErrorBoundaryProps {
message: string;
href: string;
linkLabel: string;
}
function ResourceRouteErrorBoundary({
message,
href,
linkLabel,
}: ResourceRouteErrorBoundaryProps): ReactElement {
useRouteError();
return (
<div role="alert" className="flex min-h-screen flex-col items-center justify-center gap-3 p-8">
<p className="text-sm font-medium">{message}</p>
<a href={href} className="text-sm underline">
{linkLabel}
</a>
</div>
);
}
export function ProjectsRouteErrorBoundary(): ReactElement {
return (
<ResourceRouteErrorBoundary
message="Something went wrong loading projects."
href="/projects"
linkLabel="Reload projects"
/>
);
}
export function ProjectDetailRouteErrorBoundary(): ReactElement {
return (
<ResourceRouteErrorBoundary
message="Something went wrong loading this project."
href="/projects"
linkLabel="Back to projects"
/>
);
}
export function TasksRouteErrorBoundary(): ReactElement {
return (
<ResourceRouteErrorBoundary
message="Something went wrong loading tasks."
href="/tasks"
linkLabel="Reload tasks"
/>
);
}
+144
View File
@@ -0,0 +1,144 @@
import { act } from 'react';
import { createRoot, type Root } from 'react-dom/client';
import { createMemoryRouter, RouterProvider, type RouteObject } from 'react-router-dom';
import { afterAll, afterEach, beforeAll, describe, expect, it, vi } from 'vitest';
import { taskFixtures } from './page-fixtures';
const { apiMock } = vi.hoisted(() => ({
apiMock: vi.fn(),
}));
vi.mock('@/lib/api', () => ({
api: apiMock,
}));
import { TasksPage } from './tasks';
interface Deferred<T> {
promise: Promise<T>;
resolve: (value: T) => void;
}
function createDeferred<T>(): Deferred<T> {
let resolve!: (value: T) => void;
const promise = new Promise<T>((res) => {
resolve = res;
});
return { promise, resolve };
}
let root: Root | null = null;
let container: HTMLDivElement;
beforeAll(() => {
Object.defineProperty(globalThis, 'IS_REACT_ACT_ENVIRONMENT', {
configurable: true,
value: true,
});
});
afterAll(() => {
Reflect.deleteProperty(globalThis, 'IS_REACT_ACT_ENVIRONMENT');
});
afterEach(async () => {
await act(async () => {
root?.unmount();
});
document.body.replaceChildren();
root = null;
apiMock.mockReset();
});
async function renderTasksPage(): Promise<void> {
const routes: RouteObject[] = [{ path: '/tasks', element: <TasksPage /> }];
const router = createMemoryRouter(routes, { initialEntries: ['/tasks'] });
container = document.createElement('div');
document.body.append(container);
root = createRoot(container);
await act(async () => {
root?.render(<RouterProvider router={router} />);
});
}
function clickButtonByText(text: string): void {
const button = [...container.querySelectorAll('button')].find((candidate) =>
candidate.textContent?.includes(text),
);
if (!button) {
throw new Error(`Button containing "${text}" not found`);
}
button.dispatchEvent(new MouseEvent('click', { bubbles: true }));
}
describe('TasksPage', () => {
it('shows a visible loading state before the tasks request settles', async () => {
const deferred = createDeferred<typeof taskFixtures>();
apiMock.mockReturnValueOnce(deferred.promise);
await renderTasksPage();
expect(container.textContent).toContain('Loading tasks...');
await act(async () => {
deferred.resolve(taskFixtures);
await deferred.promise;
});
});
it('starts in kanban view, toggles to list view, and opens the read-only modal from cards and rows', async () => {
apiMock.mockResolvedValueOnce(taskFixtures);
await renderTasksPage();
expect(container.textContent).toContain('Not Started');
expect(container.textContent).toContain('In Progress');
expect(container.textContent).toContain('Blocked');
const kanbanCard = [...container.querySelectorAll('button')].find((candidate) =>
candidate.textContent?.includes('Route /projects/:id'),
);
expect(kanbanCard).toBeTruthy();
await act(async () => {
kanbanCard?.dispatchEvent(new MouseEvent('click', { bubbles: true }));
});
expect(container.querySelector('[role="dialog"]')).toBeTruthy();
await act(async () => {
container
.querySelector('button[aria-label="Close task details"]')
?.dispatchEvent(new MouseEvent('click', { bubbles: true }));
});
expect(container.querySelector('[role="dialog"]')).toBeNull();
await act(async () => {
clickButtonByText('List');
});
const row = [...container.querySelectorAll('tr')].find((candidate) =>
candidate.textContent?.includes('Route /tasks'),
);
expect(row).toBeTruthy();
await act(async () => {
row?.dispatchEvent(new MouseEvent('click', { bubbles: true }));
});
expect(container.querySelector('[role="dialog"]')).toBeTruthy();
expect(container.textContent).toContain('Wire list and kanban modal interactions');
});
it('renders a visible alert when the tasks request fails', async () => {
apiMock.mockRejectedValueOnce(new Error('Tasks request failed'));
await renderTasksPage();
const alert = container.querySelector('[role="alert"]');
expect(alert).toBeTruthy();
expect(alert?.textContent).toContain('Tasks request failed');
});
});
+92
View File
@@ -0,0 +1,92 @@
import { useEffect, useState, type ReactElement } from 'react';
import { KanbanBoard } from '@/components/tasks/kanban-board';
import { TaskDetailModal } from '@/components/tasks/task-detail-modal';
import { TaskListView } from '@/components/tasks/task-list-view';
import { api } from '@/lib/api';
import { cn } from '@/lib/cn';
import type { Task } from '@/lib/types';
import { getErrorMessage } from './page-errors';
type ViewMode = 'list' | 'kanban';
export function TasksPage(): ReactElement {
const [tasks, setTasks] = useState<Task[]>([]);
const [view, setView] = useState<ViewMode>('kanban');
const [loading, setLoading] = useState(true);
const [error, setError] = useState<string | null>(null);
const [selectedTask, setSelectedTask] = useState<Task | null>(null);
useEffect(() => {
let cancelled = false;
void api<Task[]>('/api/tasks')
.then((response) => {
if (cancelled) return;
setTasks(response);
})
.catch((caught: unknown) => {
if (cancelled) return;
setError(getErrorMessage(caught, 'Failed to load tasks.'));
})
.finally(() => {
if (cancelled) return;
setLoading(false);
});
return () => {
cancelled = true;
};
}, []);
return (
<div className="flex min-h-screen flex-col px-4 py-6 sm:px-6">
<header className="mb-6 flex items-center justify-between gap-4 border-b px-1 pb-3">
<h1 className="text-2xl font-semibold">Tasks</h1>
<div className="flex rounded-lg border border-surface-border">
<button
type="button"
onClick={() => setView('list')}
className={cn(
'px-3 py-1.5 text-xs transition-colors',
view === 'list'
? 'bg-surface-elevated text-text-primary'
: 'text-text-muted hover:text-text-secondary',
)}
>
List
</button>
<button
type="button"
onClick={() => setView('kanban')}
className={cn(
'px-3 py-1.5 text-xs transition-colors',
view === 'kanban'
? 'bg-surface-elevated text-text-primary'
: 'text-text-muted hover:text-text-secondary',
)}
>
Kanban
</button>
</div>
</header>
{error ? (
<div role="alert" className="mb-6 rounded-lg border border-error/40 px-4 py-3 text-sm">
{error}
</div>
) : null}
{loading ? (
<p className="py-8 text-center text-sm text-text-muted">Loading tasks...</p>
) : view === 'kanban' ? (
<KanbanBoard tasks={tasks} onTaskClick={setSelectedTask} />
) : (
<TaskListView tasks={tasks} onTaskClick={setSelectedTask} />
)}
{selectedTask ? (
<TaskDetailModal task={selectedTask} onClose={() => setSelectedTask(null)} />
) : null}
</div>
);
}
+39
View File
@@ -3,6 +3,10 @@ import { describe, expect, it } from 'vitest';
import type { RouteObject } from 'react-router-dom'; import type { RouteObject } from 'react-router-dom';
import { routes } from '@/routes'; import { routes } from '@/routes';
import { Placeholder } from '@/spa/placeholder'; import { Placeholder } from '@/spa/placeholder';
import { ChatPage } from '@/spa/pages/chat';
import { ProjectDetailPage } from '@/spa/pages/project-detail';
import { ProjectsPage } from '@/spa/pages/projects';
import { TasksPage } from '@/spa/pages/tasks';
function collectPaths(routeObjects: RouteObject[]): string[] { function collectPaths(routeObjects: RouteObject[]): string[] {
return routeObjects.flatMap((route) => [ return routeObjects.flatMap((route) => [
@@ -55,4 +59,39 @@ describe('SPA route table', () => {
expect(element.type).not.toBe(Placeholder); expect(element.type).not.toBe(Placeholder);
}, },
); );
it('renders the real chat page instead of the P1 placeholder at /chat, inside the authenticated group', () => {
const authPaths = collectPaths(routes.at(1)?.children ?? []);
expect(authPaths).toContain('/chat');
const element = findRoute(routes, '/chat')?.element;
expect(isValidElement(element)).toBe(true);
if (!isValidElement(element)) throw new Error('Missing route element for /chat');
expect(element.type).not.toBe(Placeholder);
expect(element.type).toBe(ChatPage);
});
it.each([
['/projects', ProjectsPage],
['/projects/:id', ProjectDetailPage],
['/tasks', TasksPage],
])(
'renders a real authenticated page instead of the P1 placeholder at %s',
(path, expectedType) => {
const element = findRoute(routes, path)?.element;
expect(isValidElement(element)).toBe(true);
if (!isValidElement(element)) throw new Error(`Missing route element for ${path}`);
expect(element.type).not.toBe(Placeholder);
expect(element.type).toBe(expectedType);
},
);
it.each(['/chat', '/projects', '/projects/:id', '/tasks'])(
'defines an error boundary for %s',
(path) => {
const route = findRoute(routes, path);
expect(route?.errorElement).toBeTruthy();
expect(isValidElement(route?.errorElement)).toBe(true);
},
);
}); });
+1
View File
@@ -16,6 +16,7 @@ export default defineConfig({
}, },
server: { server: {
port: 3100, port: 3100,
strictPort: true,
proxy: { proxy: {
'/api': gatewayTarget, '/api': gatewayTarget,
'/socket.io': { target: gatewayTarget, ws: true }, '/socket.io': { target: gatewayTarget, ws: true },
@@ -0,0 +1,33 @@
# CI Queue Guard Purpose Semantics
- **Issue:** #1146
- **Target branch:** `next`
## Problem
`ci-queue-wait.sh` treats any result other than terminal success as asserted non-readiness. That is correct for merge readiness, but incorrect for the pre-push queue guard: a terminal failure or an empty status set means no pipeline is queued or running, so the queue is clear.
## Design
Make final-state handling purpose-sensitive while preserving the existing provider and payload safeguards:
- `--purpose push`
- wait while state is `pending`;
- return success for `terminal-success`, `terminal-failure`, and `no-status`;
- continue rejecting `malformed`, `unknown`, and unrecognized states.
- `--purpose merge`
- return success only for `terminal-success`;
- continue rejecting `terminal-failure`, `no-status`, malformed, unknown, and unrecognized states.
- `--require-status` remains authoritative: `no-status` fails for either purpose when it is supplied.
Diagnostics will explicitly distinguish a queue-clear push result from successful CI so callers cannot mistake an old failure for a green pipeline.
## Testing
Extend the process-level tri-state regression harness with separate push and merge assertions:
1. Push passes for terminal success, terminal failure, and no status.
2. Push still fails for pending, malformed, and unknown states.
3. `--require-status` makes push/no-status fail.
4. Merge behavior remains fail-closed except for terminal success.
5. Existing provider-unavailable audit behavior remains unchanged.
@@ -0,0 +1,208 @@
# CI Queue Guard Purpose Semantics Implementation Plan
> **For Claude:** REQUIRED SUB-SKILL: Use superpowers:executing-plans to implement this plan task-by-task.
**Goal:** Make the pre-push CI queue guard pass when no pipeline is queued or running while preserving fail-closed merge readiness.
**Architecture:** Keep provider lookup and tri-state classification unchanged. Make only the final state dispatch purpose-sensitive: push treats valid non-pending states as queue-clear, while merge continues to require terminal success. Preserve `--require-status`, malformed-payload rejection, unknown-state rejection, and audited provider-unavailable behavior.
**Tech Stack:** Bash, process-level shell regression harnesses, Gitea/GitHub status APIs.
---
### Task 1: Freeze Purpose-Specific State Semantics
**Files:**
- Modify: `packages/mosaic/framework/tools/git/test-ci-queue-wait-tristate.sh`
- Test: `packages/mosaic/framework/tools/git/test-ci-queue-wait-tristate.sh`
**Step 1: Add failing push assertions**
Change push expectations so `terminal-failure` and `no-status` require exit 0 plus an explicit `queue-clear` diagnostic. Add a `--require-status` assertion that keeps push/no-status non-zero.
**Step 2: Add failing merge assertions**
Invoke the same harness with `MOSAIC_TEST_PURPOSE=merge` and assert terminal failure and no status remain non-zero while terminal success remains zero.
**Step 3: Add unknown-state coverage**
Add a stub payload with a syntactically valid but unsupported status value and assert both purposes reject it.
**Step 4: Run the focused test and verify RED**
Run:
```bash
bash packages/mosaic/framework/tools/git/test-ci-queue-wait-tristate.sh
```
Expected: failures showing push terminal-failure and no-status returned exit 3 instead of exit 0 or lacked `queue-clear` diagnostics.
**Step 5: Commit the failing tests**
```bash
git add packages/mosaic/framework/tools/git/test-ci-queue-wait-tristate.sh
git commit -m "test(ci): define purpose-aware queue readiness"
```
### Task 2: Implement Purpose-Sensitive Final-State Dispatch
**Files:**
- Modify: `packages/mosaic/framework/tools/git/ci-queue-wait.sh:458-481`
- Test: `packages/mosaic/framework/tools/git/test-ci-queue-wait-tristate.sh`
- Test: `packages/mosaic/framework/tools/git/test-ci-queue-wait-github-checks.sh`
**Step 1: Implement push queue-clear behavior**
For `no-status`, retain the existing `--require-status` failure. Otherwise, return success for push with an explicit diagnostic such as:
```text
[ci-queue-wait] queue-clear state=no-status purpose=push branch=<branch>; no queued or running CI.
```
For `terminal-failure`, return success only for push with the same queue-clear wording. Merge must continue returning asserted non-readiness.
**Step 2: Preserve malformed and unknown rejection**
Keep `malformed`, `unknown`, and unrecognized states non-zero for both purposes.
**Step 3: Run focused tests and verify GREEN**
Run:
```bash
bash packages/mosaic/framework/tools/git/test-ci-queue-wait-tristate.sh
bash packages/mosaic/framework/tools/git/test-ci-queue-wait-github-checks.sh
```
Expected: both scripts exit 0 and report their regression suites passed.
**Step 4: Commit implementation**
```bash
git add packages/mosaic/framework/tools/git/ci-queue-wait.sh
git commit -m "fix(ci): separate push queue clearance from merge readiness"
```
### Task 3: Verify, Review, and Document Evidence
**Files:**
- Modify: `docs/scratchpads/1146-ci-queue-purpose.md`
**Step 1: Run shell syntax and focused regressions**
```bash
bash -n packages/mosaic/framework/tools/git/ci-queue-wait.sh
bash -n packages/mosaic/framework/tools/git/test-ci-queue-wait-tristate.sh
bash packages/mosaic/framework/tools/git/test-ci-queue-wait-tristate.sh
bash packages/mosaic/framework/tools/git/test-ci-queue-wait-github-checks.sh
```
**Step 2: Run repository quality gates**
```bash
pnpm preflight
pnpm typecheck
pnpm lint
pnpm test
pnpm format:check
```
Expected: every command exits 0.
**Step 3: Obtain independent review**
Request review of the exact branch head. Remediate all blocking findings and rerun focused and baseline gates.
**Step 4: Record evidence and commit**
Update the scratchpad with test output, review result, and residual risk, then commit it:
```bash
git add docs/scratchpads/1146-ci-queue-purpose.md
git commit -m "docs(ci): record queue guard verification"
```
### Task 4: Keep the Merge Wrapper Aligned with the `next` Lane
**Files:**
- Modify: `packages/mosaic/framework/tools/git/pr-merge.sh:97-101`
- Test: `packages/mosaic/framework/tools/git/test-pr-merge-head-pin.sh`
**Step 1: Write the failing regression**
Run the exact-head merge regression with its Gitea fixture targeting `next` and confirm the current wrapper rejects it because it only permits `main`.
**Step 2: Allow only documented integration targets**
Permit `main` and `next`; reject every other target. Do not alter exact-head pinning, queue-guard invocation, provider selection, or merge method enforcement.
**Step 3: Run focused merge regressions**
```bash
bash packages/mosaic/framework/tools/git/test-pr-merge-head-pin.sh
bash packages/mosaic/framework/tools/git/test-pr-merge-queue-branch.sh
bash packages/mosaic/framework/tools/git/test-pr-merge-gitea-empty-uid.sh
```
Expected: all pass, including a Gitea merge fixture targeting `next`.
**Step 4: Commit**
```bash
git add packages/mosaic/framework/tools/git/pr-merge.sh packages/mosaic/framework/tools/git/test-pr-merge-head-pin.sh
git commit -m "fix(ci): allow reviewed merges into next"
```
### Task 5: Activate and Deliver Through `next`
**Files:**
- Installed output: `~/.config/mosaic/tools/git/ci-queue-wait.sh`
**Step 1: Activate through the canonical installer**
From the reviewed worktree, run the framework installer in sync-only keep mode so operator files remain protected:
```bash
MOSAIC_SYNC_ONLY=1 MOSAIC_INSTALL_MODE=keep MOSAIC_SKIP_SKILLS_SYNC=1 \
bash packages/mosaic/framework/install.sh
```
**Step 2: Verify installed/source parity**
```bash
cmp -s \
packages/mosaic/framework/tools/git/ci-queue-wait.sh \
~/.config/mosaic/tools/git/ci-queue-wait.sh
```
Expected: exit 0.
**Step 3: Run mandatory pre-push queue guard**
```bash
~/.config/mosaic/tools/git/ci-queue-wait.sh --purpose push -B fix/1146-ci-queue-purpose
```
Expected: branch-absent or queue-clear success.
**Step 4: Push and open a PR against `next`**
```bash
git push -u origin fix/1146-ci-queue-purpose
~/.config/mosaic/tools/git/pr-create.sh \
-t "fix(ci): make queue guard purpose-sensitive" \
-b "Closes #1146" \
-B next \
-H fix/1146-ci-queue-purpose \
-i 1146
```
**Step 5: Complete reviewed integration**
Wait for exact-head terminal-green CI, obtain the required review, merge via the Mosaic wrapper, verify merged CI, and close #1146. Do not bypass any gate.
@@ -0,0 +1,71 @@
# #1019 — Zero-timeout queue-guard harness race
- **Issue:** #1019 (parent status remains `believed-fixed, pending jarvis validation`; do not close)
- **Branch:** `fix/1019-ci-queue-timeout-harness`
- **Owner:** `be-coder-08`
- **Base:** `origin/main` at `5916aeefd6ed12bcac086c6834c7f6c4ae38e1bc`
- **Charter:** `/home/hermes/agent-work/tl-mosaic/CHARTER-1019-HARNESS-FIX.md`
## Objective
Make `test-ci-queue-wait-tristate.sh` deterministic without changing any asserted outcome. Remove the indiscriminate zero-timeout race, require every status-classification case to prove the provider was observed, and prove the harness-controlled virtual clock is active.
## Scope
- In scope: `packages/mosaic/framework/tools/git/test-ci-queue-wait-tristate.sh` only, plus this evidence scratchpad.
- Out of scope: guard parsers, D2/D3 behavior, installer/reseed staleness, PR #1060, and issue closure.
## Acceptance criteria
1. RED deterministically reproduces deadline pre-emption before the provider call.
2. Every case that intends status classification positively proves provider observation.
3. Pending observes `pending` before deterministic virtual-time expiration.
4. The virtual clock has a positive interception control; a broken-clock mutant makes the suite red.
5. The exact CI-base image passes the final harness repeatedly with zero failures.
6. Baseline gates, independent code/security review, exact-head CI, and coordinator-authorized squash merge pass.
## Plan
1. Add deterministic RED instrumentation for the known merge/provider-unreachable pre-emption.
2. Replace global `-t 0` with a nonzero timeout interpreted under an event-driven virtual clock; stub sleep without wall waiting.
3. Add provider-observation and virtual-clock positive controls without changing outcome assertions.
4. Run focused shell checks, repeat in exact CI-base image, baseline gates, and independent reviews.
5. Commit with both identity layers, queue-guard plus direct Woodpecker terminal-state verification, push, self-post PR, verify poster/head/CI, obtain coordinator merge authorization, then squash merge without closing #1019.
## Budget
- No explicit token cap supplied. Keep scope to one harness file and one scratchpad; stop/report at the charter's 60% context gate.
## Evidence
- RED, deterministic pre-provider expiry: `evidence/1019-harness-fix/red-pre-provider-expiry.log` — rc 1; merge/provider-unreachable got rc 124 instead of 75, omitted CANNOT_ASSERT, did not observe the status provider, and wrote no additional audit record (four named failures).
- GREEN host focused harness: `evidence/1019-harness-fix/green-host.log` — rc 0, all outcome classes passed.
- Load-bearing clock negative control: a temporary same-directory mutant replaced the virtual `date` body with `/bin/date`; `evidence/1019-harness-fix/red-clock-not-intercepted.log` — rc 1 with named `virtual clock interception did not run` failures. The mutant file was removed after the run.
- Exact CI-base repeat: `git.mosaicstack.dev/mosaicstack/stack/ci-base:latest`, repository mounted read-only, harness work under container `/tmp`; `evidence/1019-harness-fix/ci-image-repeat/summary.log`**100 pass / 0 fail / 100 total**.
- Synchronization design: provider-status observation creates the event marker; virtual time is 1000 before the event and 1002 afterward. Pending alone reaches the stubbed no-op sleep and a post-observation deadline check. `-t 1` is uniquely load-bearing because removing it restores the 900-second default deadline at virtual time 1900, which 1002 does not cross. The numeric timeout is subject semantics under virtual time, not a wall-clock synchronization duration.
## Review remediation — semantic timeout vs. liveness bound
Security review found that virtual time remained at 1000 forever before provider observation and stubbed sleep never waited. A regression looping before the status endpoint—or blocking in the first provider call—therefore could prevent `run_guard` from returning, so the post-return provider assertion could never fire.
**General rule:** A timeout usually serves two purposes: semantics and liveness. Removing wall time from semantic synchronization can silently remove the only independent hang bound. Preserve deterministic virtual time for subject semantics, but provide a separately implemented real-clock liveness watchdog and prove that watchdog fires.
Remediation:
- Every guard subject invocation is launched by absolute `/usr/bin/python3` in a new session. Python's internal monotonic `wait(timeout=...)` provides real-clock liveness independently of PATH; expiry kills the entire isolated process group, so neither PATH-front shims nor a blocked provider descendant can retain the capture pipe.
- Watchdog expiry returns distinct harness rc 90 plus `FAIL HANG watchdog`, separate from subject timeout rc 124.
- A first attempt using absolute `/usr/bin/timeout -s KILL` passed on GNU coreutils but failed in the exact Alpine CI-base image: BusyBox killed the immediate wrapper while the guard/provider descendants survived and retained the command-substitution pipe. The process-group kill is therefore required behavior, not portability polish.
- A committed positive control hangs the branch-provider stub before the status endpoint. It must terminate through the watchdog, emit the hang-specific diagnostic, return rc 90, and prove the status provider was never reached.
- RED before remediation: a temporary ordinary-success mutant hung before provider observation; only an external control could kill the suite (rc 137), and there was no internal hang-specific diagnostic (`red-watchdog-absent.log`).
- The watchdog mutant/control is load-bearing: removing the internal watchdog leaves the control unable to produce its required rc 90 and diagnostic.
Post-review evidence:
- Host focused harness with process-group watchdog: rc 0 (`green-watchdog-process-group-host.log`).
- Exact Alpine CI-base focused harness with process-group watchdog: rc 0 (`green-watchdog-ci-image.log`).
- Hanging ordinary-success mutant: suite rc 1; success returned rc 90, emitted `FAIL HANG watchdog`, and loudly reported that provider/clock observation did not occur (`red-watchdog-fires.log`).
- Removed-`-t 1` mutant: suite rc 1; pending was terminated by the watchdog instead of producing `ASSERTED_NOT_READY`, proving the explicit timeout is load-bearing (`red-timeout-argument-removed.log`).
## 60% context hold
Stopped before baseline/review/commit as required by the charter. Remaining: inspect final diff, shell/static/baseline gates, independent code/security review, remediation if any, identity-bound commit/trailer verification, mandatory queue guard plus direct terminal Woodpecker `mosaic` enumeration, push, self-posted PR/provider poster read-back, exact-head terminal-green CI, coordinator merge authorization, squash merge, main CI verification, and leave #1019 unclosed as `believed-fixed, pending jarvis validation`.
+65
View File
@@ -0,0 +1,65 @@
# #1146 — CI Queue Guard Purpose Semantics
## Objective
Make the pre-push queue guard wait for queued/running CI without requiring the previous remote head to have successful CI. Preserve fail-closed merge readiness.
## Scope
- `packages/mosaic/framework/tools/git/ci-queue-wait.sh`
- focused queue-guard regression tests
- design and scratchpad documentation
- local framework activation required before the fixed guard can authorize this branch's push
## Plan
1. Freeze purpose-specific behavior in failing process-level tests.
2. Implement the smallest state-dispatch change.
3. Run focused shell tests and repository quality gates.
4. Obtain independent review and remediate findings.
5. Install the reviewed framework source locally, run the mandatory pre-push queue guard, and push.
6. Open a PR against `next`, verify terminal-green CI, and close #1146 after merge.
## Budget
- ASSUMPTION: no explicit token cap was provided.
- Working estimate: 12K tokens.
- Scope reduction: change only final-state dispatch and focused tests; do not redesign provider adapters.
## Progress
- Confirmed source and installed guards are byte-identical.
- Reproduced `terminal-failure` blocking `--purpose push`.
- Root cause: final-state dispatch requires terminal success for both push and merge.
- Design approved: push is queue-clear on valid non-pending states; merge remains fail-closed.
## Tests
- RED confirmed before implementation: the focused tri-state harness reported push `terminal-failure` and `no-status` as `ASSERTED_NOT_READY`.
- GREEN: `bash packages/mosaic/framework/tools/git/test-ci-queue-wait-tristate.sh` — all outcome classes passed.
- GREEN: `bash packages/mosaic/framework/tools/git/test-ci-queue-wait-github-checks.sh` — 6/6 purpose-aware cases passed.
- GREEN: `bash -n` passed for the changed guard and both focused harnesses.
- GREEN: `pnpm preflight`, `pnpm typecheck`, and `pnpm lint` passed.
- `pnpm test` ran 45/46 workspace test tasks successfully, but the pre-existing Gateway `cross-user-isolation.test.ts` failed during cleanup with PostgreSQL error `28P01` (local `mosaic` password authentication failure). The changed Mosaic framework test task passed within that run.
- GREEN: focused queue and merge shell regressions passed after the wrapper change.
- GREEN: isolated Mosaic Vitest run passed (81 files, 1,514 tests).
- The normal parallel Mosaic Vitest run has an environment-sensitive pre-existing failure in `install-ordering-guard.spec.ts`: the real activation probe changes between two calls while other suites run concurrently. Running the same spec alone and the complete Vitest suite with one fork passes.
- The framework shell suite's pre-existing `version_coupling_unittest.py` also fails locally because the newly installed `mosaic` is now on PATH despite the test injecting a nonexistent PATH; CI's clean image does not have this global CLI. All changed queue/merge harnesses pass.
- GREEN: `pnpm format:check` passed.
- Note: an additional ad hoc Prettier command was not applicable to shell files because Prettier has no shell parser; the repository-wide format check passed using its configured file globs.
## Review
- Independent Codex review of the six-file diff: approved, confidence 0.84, zero blockers/should-fix/suggestions.
- Review confirmed push queue-clear behavior, merge fail-closed behavior, and `--require-status` coverage.
## Risks and Blockers
- Canonical framework activation completed with `MOSAIC_SYNC_ONLY=1 MOSAIC_INSTALL_MODE=keep MOSAIC_SKIP_SKILLS_SYNC=1 bash packages/mosaic/framework/install.sh`.
- Source and installed queue guards are byte-identical (`cmp` and SHA-256 parity passed).
- The installed pre-push guard now passes for the not-yet-remote feature branch with `queue clear`.
- The required merge wrapper then exposed a second bootstrap defect: `pr-merge.sh` hardcoded `main`, contradicting the documented PR-based `next` integration lane. Tracked as #1149 and fixed in the same delivery branch with a regression fixture targeting `next`.
- Activation emitted the existing manifest-safety warning that six `fleet/run/*.hb*` operator files were touched then restored; no data loss was observed, but this remains a pre-existing framework-manifest defect to report separately.
- The first activation attempt timed out after 600 seconds while copying the 113K-file operator snapshot; the bounded 1,800-second retry completed successfully. It left a partial durable snapshot from the interrupted attempt in the normal backup directory; the completed snapshot is the newer `pre-update-20260810T195317Z` entry.
- Full baseline test completion is blocked by the unrelated local PostgreSQL authentication/cleanup failure described above; CI has its own disposable PostgreSQL service.
- Existing `.mosaic/orchestrator/*` working-tree changes are unrelated and must remain unstaged.
@@ -0,0 +1,93 @@
# W-B — Measure Pi's real tool registry
- **Task / internal ref:** W-B from the lease-remediation orchestrator brief (no matching `docs/TASKS.md` row; workers do not modify that file)
- **Objective:** identify the exact tool names emitted as `event.toolName` by the installed Pi runtime and compare them with the broker's Pi read-only carve-out.
- **Scope:** measurement and report only; no broker or runtime source changes. W-C is out of scope.
- **Budget:** no explicit token cap; constrained to this scratchpad and one local commit.
- **Installed runtime:** `@earendil-works/pi-coding-agent` / `pi` `0.84.1`.
## Method
I created a throwaway extension at `/tmp/measure-pi-tool-registry.ts` (not in the worktree). On `session_start` it recorded `pi.getAllTools()` and `pi.getActiveTools()`; on every `tool_call` it appended the exact `event.toolName`. I then launched an isolated, ephemeral Pi session with all built-ins explicitly selected:
```text
PI_OFFLINE=1 pi --mode print --no-session --no-approve \
--no-context-files --no-skills --no-prompt-templates --no-extensions \
-e /tmp/measure-pi-tool-registry.ts \
--tools read,bash,edit,write,grep,find,ls <deterministic probe prompt>
```
The prompt exercised file read, content search, file search, directory listing, shell execution, file write, and file edit. Pi exited `0`; every selected tool produced one `tool_call`. The write/edit control artifact ended with exact content `after`, proving the mutating calls executed in order.
This runtime observation was cross-checked against the installed distribution's canonical registry at `dist/core/tools/index.js:17`, which declares the same seven names. The gate consumes the measured field directly at `packages/mosaic/framework/runtime/pi/mosaic-extension.ts:368`.
## Exact distinct built-in set
The installed Pi built-in registry is exactly:
```text
{bash, edit, find, grep, ls, read, write}
```
| Tool | Runtime registry observation | `tool_call` observation | Installed definition |
| --- | --- | --- | --- |
| `read` | `<builtin:read>` | observed once | `dist/core/tools/read.js:138` |
| `bash` | `<builtin:bash>` | observed once | `dist/core/tools/bash.js:231` |
| `edit` | `<builtin:edit>` | observed once | `dist/core/tools/edit.js:170` |
| `write` | `<builtin:write>` | observed once | `dist/core/tools/write.js:138` |
| `grep` | `<builtin:grep>` | observed once | `dist/core/tools/grep.js:79` |
| `find` | `<builtin:find>` | observed once | `dist/core/tools/find.js:79` |
| `ls` | `<builtin:ls>` | observed once | `dist/core/tools/ls.js:61` |
The raw distinct `event.toolName` result was:
```json
["bash", "edit", "find", "grep", "ls", "read", "write"]
```
Pi registers all seven, but its default active set is only `read`, `bash`, `edit`, and `write` (`dist/core/sdk.js:132`). The probe explicitly activated all seven so the three search/list tools could be observed at the hook.
## Positive control
The known `read` tool was the control. The method surfaced it twice:
1. `pi.getAllTools()` returned `read` with source path `<builtin:read>`.
2. Reading `/tmp/pi-registry-probe/seed.txt`, which contained `CONTROL_TOKEN`, produced one hook record with `event.toolName === "read"`.
The control was therefore positive; the seven-name result is measured, not an empty-probe inference.
## Carve-out comparison and collision result
The broker currently declares `{"read", "grep", "find", "ls"}` at `packages/mosaic/framework/tools/lease-broker/daemon.py:54`.
- `read`: real built-in.
- `grep`: real built-in.
- `find`: real built-in.
- `ls`: real built-in.
All four carve-out names are exact, case-sensitive Pi tool names.
The general execution/writing tool names are `bash`, `edit`, and `write`. Their intersection with the carve-out is empty:
```text
{bash, edit, write} ∩ {read, grep, find, ls} = ∅
```
Therefore no general shell-exec or file-mutating Pi tool shares a name with a carve-out entry. `grep` and `find` may invoke constrained search helpers internally, but neither exposes an arbitrary command interface; the arbitrary command tool is distinctly named `bash`.
The Mosaic extension separately registers the non-built-in custom tool `mosaic_context_recover` at `packages/mosaic/framework/runtime/pi/mosaic-extension.ts:379`; the broker handles that identity through its dedicated recovery exemption rather than the read-only set (`daemon.py:722`). Unknown or third-party custom tools are not part of Pi's built-in seven-name registry and remain outside the carve-out.
## Verification evidence
- `pi --version``0.84.1`.
- Isolated probe exit → `0`.
- Runtime `getAllTools()` count → `7`, all with `sourceInfo.source === "builtin"`.
- Distinct hook names → `bash`, `edit`, `find`, `grep`, `ls`, `read`, `write`.
- Hook counts → exactly one call for each of the seven names.
- Mutation artifact after `write` then `edit` → exact content `after`.
- Installed registry source → `allToolNames = new Set(["read", "bash", "edit", "write", "grep", "find", "ls"])`.
## Risks / limitations
- The probe deliberately disabled all other extensions, so extension-defined third-party tools were excluded from the built-in registry measurement. The production gate still receives those names and treats names outside the broker carve-out as mutating/fail-closed.
- Explicit `--tools` activation was required to exercise `grep`, `find`, and `ls`; this does not imply they are active in Pi's default four-tool configuration.
@@ -0,0 +1,99 @@
# PR merge squash message field
- **Charter:** `/home/hermes/agent-work/CHARTER-PRMERGE-MESSAGE-FIELD.md`
- **Owner:** `be-coder-08`
- **Branch:** `fix/pr-merge-message-field`
- **Base:** remote `main` / local `origin/main` at `85d2108e4ed15c744ad3b87a5b629e7b2d39405a`
- **Estate:** HOMELAB tooling shared by HOMELAB and USC
## Objective
Add an optional, identity-checked Gitea squash message to `pr-merge.sh` so genuine multi-author PRs retain non-poster branch authors without weakening hardcoded squash behavior.
## Binding requirements
1. `Do` remains hardcoded to `squash`; no provider/repository default may select merge style.
2. A verified trailer uses a PR commit's linked `author.login` and that same commit's author email. No `/users/{login}` primary-email lookup occurs. Recorded rationale: this asks only what the provider can answer.
3. A commit with `author.login` null blocks before merge, prints both the null provider fact and commit email fact, and names the escalation principal.
4. The BLOCK arm must be observed firing; a normal canonical single-author API payload remains explicit squash plus its reviewed `head_commit_id`.
5. Every provider mutation is read back from the provider; no real PR is merged during tests.
## Derived interface decisions
- Add `--co-author-trailers` rather than accepting arbitrary message text. The wrapper enumerates PR commits and constructs trailers, making an unchecked `Co-authored-by` line unexpressible.
- Require `--escalate-to PRINCIPAL` with `--co-author-trailers`, so the BLOCK diagnostic always names a principal rather than a generic role.
- Do not expose `MergeTitleField` separately. When trailers exist, set it from the provider PR title and set `MergeMessageField` only to construction-generated trailers. This preserves one provider source for the title and avoids an unrelated caller-controlled degree of freedom.
- Preserve first-commit order and emit one trailer per distinct non-poster `author.login`, using that first linked commit's own email.
## Canonical delivery plan
1. Port the capability into the installed source of truth, `packages/mosaic/framework/tools/git/pr-merge.sh`; do not retain `infra/fleet/tools/git` as a second copy.
2. Preserve canonical `--expect-head`, exact head branch/repository/SHA queue inspection, Gitea atomic head pinning, GitHub `--match-head-commit`, and delete-after-merge semantics.
3. Do not port the deployed-only `--skip-queue-guard` bypass. Add the focused harness to the canonical framework-shell suite and re-establish RED/GREEN on the packaged baseline.
4. Deliver through a reviewed package release followed by `mosaic update` with its default framework reseed. The installer snapshots, manifest-syncs framework-owned `tools/**`, and rolls back on failure.
5. Before either estate relies on the change, require installed/package hash equality, `MergeMessageField` presence, and a green focused harness. Release/reseed ownership is currently unassigned and blocks activation after source merge.
## Evidence
- RED against the byte-identical deployed baseline (`sha256 08a65e8584c5…`): rc 1 with eight named failures. The wrapper rejected `--co-author-trailers`; the null-login path emitted none of the required BLOCK facts/principal; and both verified/ordinary API paths failed the stdin-config credential assertion (ordinary path exposed the fixture token through curl argv). Log: `/home/hermes/agent-work/be-coder-08/evidence/prmerge-message-field-red.log`.
- GREEN on the deployed-baseline candidate: verified linked multi-author payload, null-login BLOCK, required named principal, explicit squash, stdin-config token transport, and absence of `/users` lookup all passed. Log: `/home/hermes/agent-work/be-coder-08/evidence/prmerge-message-field-green.log`.
- RED against canonical packaged baseline `c581ef48…`: rc 1 with 32 assertions. It rejects the new option, and the first harness version did not satisfy canonical head branch/repository/SHA metadata. Log: `/home/hermes/agent-work/be-coder-08/evidence/prmerge-packaged-baseline-red.log`. The port adapts the fixture rather than weakening canonical head controls.
- Provider capability probe against `git.mosaicstack.dev`: authenticated `be-coder-08` POST to deliberately nonexistent PR `2147483647` with both message fields returned JSON HTTP 404; the unauthenticated same request returned JSON HTTP 401 (not the charter's predicted 403). The authenticated-vs-unauthenticated differential proves write authorization resolved while no mergeable subject existed. `tl-mosaic` ruled the literal non-load-bearing: preserve the observed 404/401 pair and do not manufacture a 403 case. No cause was inferred and no real PR was targeted.
- Provider-generated trailer behavior is not treated as exclusive or absent. The wrapper's VERIFIED/BLOCK decision binds each requested non-poster trailer to commit `author.login` plus that commit's email; it does not assume `MergeMessageField` is the squash's only trailer source. The poster is omitted from the constructed list because the resulting squash author already records the poster; any additional provider-generated trailer is outside this change's unmeasured mechanism.
- An early candidate SHA-256 `5de32876990e4f26920448cb3220cc7f1146d558b4dd2bc1ee1a2abee2f2cbe6` passed the initial harness, then author-side review found credential-fallback and argv-exposure defects. The live deployed wrapper was atomically restored to baseline SHA-256 `08a65e8584c52c6d41ea1c686f8b95585c21e4b37320a2447eba09359a0e02c1`; the remediated candidate remains only in the worktree.
## Remediation and current review state
1. Token and Basic Auth now use stdin curl configuration, not argv. PR title, contributor email, and the JSON payload also remain out of child argv.
2. Each credential attempt binds commit inspection and merge. A token failure during either inspection or mutation causes Basic fallback to repeat inspection before mutation; the payload pins the inspected `head_commit_id`.
3. Focused tests cover token-resolution fail-closed behavior, both HTTP-401 fallback seams, metadata/credential argv absence, null-login BLOCK, explicit squash, canonical reviewed-head binding, unchanged ordinary payload, and retained log-safe provider diagnostics. Token-resolution RED: `/home/hermes/agent-work/be-coder-08/evidence/prmerge-token-resolution-red.log`.
4. Codex review rounds 35 requested retained provider error text, log-safe provider diagnostics, fail-closed credential fallback, stable value-option parsing, and PR-title trailer-injection prevention. These are remediated with regression assertions. A post-remediation independent review is still required.
5. **Accepted linkage limitation:** `author.login` resolution proves that the commit address maps to a registered provider account. It does not prove that the named principal authored the commit because Git author metadata is self-asserted. This gate checks attribution linkage, not authorship; commit signing is out of scope and currently unadopted. Coordinators explicitly ruled that this does not add a third state.
6. Codex's sandbox could not execute the harness because its checkout was read-only; that environmental limitation is recorded separately from host-side test results.
## Disposable provider fixture acceptance
- Use a retained scratch repository only, with two branch authors and `author != committer` on at least one commit.
- Arm A supplies a message-field trailer for one non-poster; record whether that value lands without forcing the partial-pair result into under-specified `APPENDS`/`REPLACES` labels. Demonstrate an absence control.
- Arm B includes a registered trailer for a different non-poster on a branch commit; record whether it survives or drops. Verify identity through an existing commit whose `author.login` resolves and demonstrate an absence control.
- Parse landed trailers key-agnostically with `^[A-Za-z-]+-[Bb]y:` and record generated poster pair presence/absence plus resulting poster attribution.
- Record `/users/<login>` status and raw email only as non-gating estate telemetry. Never read `active`, `visibility`, or any profile field as an identity gate.
- Use distinct principals: poster `be-coder-08`, merger `Mos`, Arm A `be-coder-07`, and Arm B `be-coder-06`. Capture every trailer-shaped line verbatim and in order. Zero trailer lines means the generator did not fire and the run is `VOID`, not evidence that either arm dropped.
- Report the same read-back evidence to `mos-claude` on socket `default` and `tl-mosaic` on socket `mosaic-fleet`. Report values rather than mechanism inferences and stop on any poster-attribution regression.
## Fixture preflight
- Retained public repository: `mosaicstack/prmerge-trailer-fixture`; PR `#1`, posted by `be-coder-08` and reserved for merge by `Mos`.
- Existing `mosaicstack/stack` commits resolve `be-coder-07` and `be-coder-06` through `author.login`; exact addresses are `[email protected]` and `[email protected]`.
- Non-gating HOMELAB telemetry for authenticated reader `be-coder-08`: `/api/v1/users/be-coder-06` returned HTTP 200 with raw `email` value `[email protected]`.
- Provider preflight showed PR commit enumeration is newest-first. A new RED test proved that deriving `head_commit_id` from the final array element selected the wrong commit. The candidate now reads `.head.sha` from the authenticated PR endpoint before enumeration, verifies it appears in the commit set, and atomically pins that SHA in the explicit squash payload. RED: `/home/hermes/agent-work/be-coder-08/evidence/prmerge-head-order-red.log`.
- Fixture PR head `f6ba6e5105031fa21f5ff7bd8e4379d99c16e1de` has `author.login=be-coder-07`, `committer.login=be-coder-08`, and branch-message trailer `Co-authored-by: be-coder-06 <[email protected]>`.
## Fixture result
- `Mos` merged retained fixture PR `#1` through staged candidate SHA-256 `60e779a85fd13b729d859ea7c986d1e9b1641b97991611329226c1b3113ffb6e`; resulting squash commit: `3f550715d9bc716426fd355a65fe997b3a90fa7d` with one parent.
- Provider read-back: poster/commit author `be-coder-08`, committer/merger `Mos`. The run is non-void.
- Trailer-shaped lines, verbatim and in order:
1. `Co-authored-by: be-coder-07 <[email protected]>`
2. `Co-authored-by: be-coder-08 <[email protected]>`
- Arm A supplied field value (`be-coder-07`) landed. Arm B branch trailer (`be-coder-06`) dropped. Both fabricated absence controls remained absent. No `Co-committed-by:` line landed.
- The candidate payload construction explicitly excludes the poster and supplied only the Arm A `be-coder-07` line. Therefore the landed poster line was provider-generated, not candidate-composed. The raw result supports `FIELD LANDS`, `BRANCH DROPS`, and `POSTER GENERATED`; it does not support a claim that candidate code supplied the poster. Evidence: `/home/hermes/agent-work/be-coder-08/evidence/prmerge-fixture-readback.log` and the retained provider object.
- Retained fixture PR `#2` measured the N=2 shape needed by `#1030`: supplied `be-coder-07` then `be-coder-06`; both landed in that order, followed by the provider-generated poster line. No truncation or dedup occurred at N=2. Resulting squash: `39db9d13aed0…`.
## Current hold point
PR `mosaicstack/stack#1066` is open. Its first frozen head `f4b162fa…` was terminal-green in Woodpecker `mosaic` pipeline `#2225`, but that evidence becomes stale when the canonical port moves the head. The deployed wrapper remains baseline `08a65e85…`; no manual copy will occur. Canonical port tests, commit amendment, rebase, one guarded force-with-lease, exact-head CI, and new independent review remain. Even after source merge, activation remains blocked on an assigned package-release/reseed owner and installed-byte read-back.
## Security review 96 remediation
Exact reviewed predecessor head: `1ceb11058f64dd7f4a817ceb2124f980a1c4dd23`.
RED-first focused harness produced 10 named failures: all curl calls lacked size/time/connect bounds; raw ESC email reached mutation; oversized and stalled curl failures were discarded and reached mutation; nonempty Basic output with resolver rc 91 authorized mutation.
Security remediation:
- Removed the cross-principal HTTP-401 Basic fallback. Both inspection-401 and merge-401 paths now refuse without Basic resolution or mutation; `get_gitea_basic_auth` references in the merge subject are 0.
- Applied `--max-filesize`, `--max-time`, and `--connect-timeout` to all 3/3 provider curl sites and fail closed on curl transport rc at all 3/3 sites.
- Required linked email bytes to be ASCII and printable before constructing `MergeMessageField`; guarded construction sites 1/1.
GREEN: message-field, exact-head, empty-UID/API, queue branch/repository/SHA, bash syntax, ShellCheck, and diff check pass. R7 total-removal mutants went RED: email guard 3 rows; bound switches 1 row; transport-rc guards 4 rows; HTTP-401 refusal 3 rows. R7 bound: mutants prove total removal only; explicit denominators above prove site coverage.
@@ -0,0 +1,99 @@
# P3-R1 — Routing Health Enum + `/mcp` Wiring Scratchpad
**Task:** P3 hands-on acceptance blockers #1 and #5
**Mission:** `mvp-20260312` (active)
**Branch:** `feat/webui-p3r1-routing-mcp` from `origin/next`
**Required base:** `20718b5a273d243363a4f5cbef5bbf692a805bdb`
**Tracking ref:** Direct P3-R1 author brief; no provider issue supplied; no PR or merge authorized
**Started:** 2026-08-11T14:52:52-05:00
## Objective
Fix exactly two P3 acceptance blockers:
1. Make routing consume the canonical `ProviderHealthStatus` enum, with `healthy` and `degraded` routable and `down` non-routable, at both routing decision sites.
2. Wire `McpClientModule` into `CommandsModule`, make `McpClientService` required, remove the unreachable unavailable-service branch, and prove `/mcp status` reaches the client.
Explicitly excluded: provider registry/adapters, `provider.service.ts`, `agent.service.ts`, `chat.gateway.ts`, fallback membership, task classification, selector UI, conversation resume, reload UX, WS/origin/handshake behavior, dependencies/lockfile, and `apps/web/**` changes.
## Plan
1. Confirm exact base/branch and inspect every cited source/test anchor plus all direct `CommandExecutorService` construction sites.
2. Record baseline gateway and focused routing/commands/MCP test totals.
3. Add regression tests first and run focused tests to capture expected RED failures.
4. Apply only the typed routing helper/signature changes and required MCP module/constructor/guard changes; update impossible `up` fixtures.
5. Run focused tests, all user-required verification gates, lockfile/scope/diff checks, and record counts.
6. Obtain independent spec and code/security review; remediate any findings and repeat affected gates.
7. Commit conventionally, run the pre-push queue guard, and push only the feature branch (no PR or merge).
## Budget
No explicit token cap supplied. Working soft cap: **30K tokens**, based on two bounded gateway bug fixes, focused TDD, full gateway/root verification, independent review, and branch delivery. One coding worker will execute serially; reviews will be independent and serial to avoid worktree collisions.
## Startup Evidence
- `git fetch origin` rc=0.
- `origin/next` confirmed exactly `20718b5a273d243363a4f5cbef5bbf692a805bdb`.
- Local and remote `feat/webui-p3r1-routing-mcp` were absent before creation.
- Branch creation rc=0; HEAD equals the required base.
- Harness-owned `.mosaic/orchestrator/session.lock` was already dirty and remains excluded from staging/commit.
## Baseline Evidence
- Gateway full suite: rc=0; **64 files / 693 tests passed**, 7 files / 17 tests skipped (71 files / 710 tests total).
- Routing sub-suite (`src/agent/routing`): rc=0; **3 files / 105 tests passed**.
- Commands/MCP sub-suite (`src/commands`, `src/mcp-client`): rc=0; **6 files / 76 tests passed**.
## TDD and Implementation Evidence
- Routing RED: after canonical fixture/test changes but before the service fix, focused routing run returned rc=1 with **15 failed / 91 passed (106)** because the old `up`/`ok` gates rejected `healthy` and `degraded`.
- Routing GREEN: canonical `ProviderHealthStatus` map types, one `isRoutable` helper, and both comparison sites corrected; focused routing suite passed.
- MCP behavior test now drives `/mcp status` through a required mock client and asserts the client is called, success is returned for zero servers, and the former unavailable message is absent.
- MCP wiring mutation RED used the final `Reflect.getMetadata('imports', CommandsModule)` assertion with only the production `McpClientModule` import/registration temporarily removed: rc=1, exact failure `expected [GCModule, …] to include McpClientModule`.
- MCP wiring GREEN after byte-for-byte production restoration: rc=0. Temporary mutation did not remain.
- Every direct `new CommandExecutorService(...)` test construction now supplies a non-null MCP client mock.
- Initial Codex worker launch failed rc=1 from missing OpenAI bearer authentication. First Mosaic Claude launch failed rc=1 because Distrobox-local runtime contracts were absent; retry with the supported host `MOSAIC_HOME=/home/jwoltje/.config/mosaic` succeeded.
## Review Evidence
- Independent spec review: **approve**, 0 blockers, scope OK.
- Independent code/security review: **approve**, 0 blockers, 0 critical/high security findings. It suggested an actual module-wiring assertion, which was added with valid mutation RED/GREEN evidence.
- Independent final re-review after remediation: **approve**, 0 blockers, 0 critical/high security findings, no remaining findings.
- Optional missing-provider/`undefined` test suggestion was not adopted: the brief explicitly requires the three canonical statuses (`healthy`, `degraded`, `down`) and forbids scope expansion; runtime behavior for absent keys remains `undefined` → non-routable through the required helper signature.
## Documentation Assessment
- `docs/PRD.md` already requires provider fallback/routing and MCP capability; this increment restores implementation to those existing contracts.
- No public API endpoint, payload schema, auth/permission rule, navigation, deployment procedure, or new user workflow changes. OpenAPI, endpoint index, user/admin/developer guides, and sitemap are therefore N/A for this bounded repair.
- This append-only scratchpad is the implementation, TDD, review, and verification record. Canonical docs remain in-repo; no publishing action is in scope.
## Final Verification Evidence
All required and repository-situational gates completed with rc=0:
| Gate | Result |
| --- | --- |
| `pnpm install --frozen-lockfile` | rc=0 |
| Gateway typecheck | rc=0 |
| Gateway lint | rc=0 |
| Routing focused suite | rc=0; 3 files / 106 tests |
| Commands/MCP focused suite | rc=0; 6 files / 78 tests |
| Gateway full suite | rc=0; 64 files / 696 tests passed; 7 files / 17 tests skipped |
| Gateway build | rc=0 |
| Root typecheck | rc=0; 45/45 tasks |
| Web test | rc=0; 19 files / 154 tests |
| Root lint | rc=0; 25/25 tasks |
| Root format check | rc=0 |
| `git diff --check` | rc=0 |
- Gateway suite before→after: **693→696 passing tests**; skipped remained 17 (total 710→713).
- Routing focused before→after: **105→106 passing tests**.
- Commands/MCP focused before→after: **76→78 passing tests**.
- `pnpm-lock.yaml` SHA-256 before/after frozen install: `9acaa89d213b3281e757b6edf6fdb8727176570d725b78a0de234c61a7f3c332`; diff versus `origin/next` rc=0.
- Verified no changed path under `apps/web/**`, provider service/adapters, `agent.service.ts`, `chat.gateway.ts`, or lockfile.
- Verified both `McpClientModule` production wiring lines remain and no impossible `up`/`ok` routing status checks/fixtures remain.
- Verified code/test diff SHA-256: `27b41a855084b9dd85a7bc2a79fa3251d114ee226a4f1b835e65134c25a4f5a8`.
## Delivery State
Implementation, testing, documentation assessment, and independent review are complete. Remaining authorized actions: format this final scratchpad append, create one conventional commit, run the required push queue guard, and push only `feat/webui-p3r1-routing-mcp`; no PR or merge.
+271
View File
@@ -0,0 +1,271 @@
# WebUI Phase P — P4-1 Projects + Tasks SPA Scratchpad
**Task ID:** P4-1
**Tracking ref:** Phase P RFC §6.4 / §2.4 author brief; no provider issue supplied
**Mission context:** `mvp-20260312` (active)
**Branch:** `feat/webui-p4-1` from `origin/next`
**Started:** 2026-08-11
**Role:** orchestrator-controlled author worker; `docs/TASKS.md` remains orchestrator-only and is not part of this increment
## Original tasking
Implement the bounded P4-1 SPA parity slice exactly as briefed: real authenticated `/projects`, `/projects/:id`, and `/tasks` React Router pages ported from the existing Next baseline; reuse existing project/task components; preserve relative same-origin REST access; support only the existing project/task PATCH edit flows; add route error boundaries and page tests; pin Vite strict port 3100 and legacy Next dev/start to 3101; make no gateway, package, dependency, nav-shell, chat, settings, or admin changes; verify, commit, and push only `feat/webui-p4-1` (no PR or merge). If any required REST endpoint is absent, stop as `BLOCKED:` rather than inventing a workaround.
## Objective and acceptance map
- A: `/projects` list with loading, cards, empty, and surfaced API-error states; defer MissionStatus side panel.
- B: `/projects/:id` detail with overview/tasks/missions tabs, parallel project/mission/task load, bounded project and task PATCH edits.
- C: `/tasks` list/kanban with bounded task PATCH edits.
- D: replace the three route placeholders and add route error boundaries without changing chat.
- E: use only `@/lib/types` and `@/lib/api`; relative `/api/...` REST paths only.
- F: Vite `strictPort: true`; Next dev/start pinned to 3101; no proxy/dependency changes.
- G: fixture-backed Vitest coverage for lists, states, tabs/toggles, PATCH calls, and real route elements; all named verification gates pass.
## Plan
1. Verify the required gateway REST routes and inspect the exact `origin/next` SPA, Next baseline pages, shared components, types, API helper, and P3 test conventions.
2. Record the pre-change web test count and add required page/route tests first, observing expected RED failures.
3. Port the three pages and shared route error boundary, then wire routes and bounded PATCH flows.
4. Apply only the two dev-port pin changes.
5. Run focused tests, all user-required gates, lockfile and same-origin checks, and `git diff --check`.
6. Obtain independent spec/code/security review; remediate and repeat affected gates until clear.
7. Commit conventionally, run the required pre-push queue guard, push only the feature branch, and record exact evidence here.
## Testing strategy
TDD is applied because this adds user-visible data/edit behavior. Component tests mock only the existing API boundary and exercise rendered behavior and PATCH payloads. Primary situational evidence is the required page interaction suite plus route-resolution checks; baseline evidence is typecheck, lint, full web test, build, frozen install, formatting/diff hygiene, and same-origin grep.
## Budget
No explicit token cap was supplied. Working soft cap: **50K tokens**, derived from three coupled React pages, route/error wiring, interaction tests, port config, review/remediation, and full verification. One Codex implementation worker and independent review workers will be used serially to avoid worktree collisions.
## Base evidence
- `git fetch origin` rc=0.
- `origin/next` and branch start: `e00cc475a2b1e9866bd4e2f8df80aff640c3a543`.
- Branch created: `feat/webui-p4-1` tracking `origin/next`.
- Harness-owned `.mosaic/orchestrator/session.lock` is dirty and must remain unstaged/uncommitted.
## Progress / evidence
- [x] Loaded active mission manifest, latest scratchpad, top-level tasks, PRD, orchestration/delivery/frontend/QA/documentation/review/TypeScript guides, and matching implementation skills.
- [x] Confirmed exact base SHA and created the feature branch.
- [x] Required REST endpoints verified in Gateway source: project list/detail/PATCH, task list/filter/detail/PATCH, and mission list all exist.
- [x] Scope assumption check found a blocking contradiction before source implementation.
- [ ] Pre-change web test count recorded.
- [ ] RED tests observed.
- [ ] Implementation complete.
- [ ] Independent review clear.
- [x] Required verification gates run against the unchanged web baseline; exact leak-grep expectation is independently blocked by 12 pre-existing matches.
- [x] Blocker record committed as `5ede86a5` and feature branch pushed; no PR opened and no merge performed.
## Blocker — 2026-08-11
`P4-1` is blocked because the bounded edit UX asserted by the brief does not exist at the confirmed `origin/next` base (`e00cc475`):
- `apps/web/src/components/tasks/task-detail-modal.tsx` is read-only. Its props are only `task` and `onClose`; it contains no input/select/textarea, update callback, or `api()`/PATCH call.
- `apps/web/src/app/(dashboard)/projects/[id]/page.tsx` contains no project edit controls and no project/task PATCH call.
- `apps/web/src/app/(dashboard)/tasks/page.tsx` does not open `TaskDetailModal`; its click handler only logs that a future detail view will be added.
- A repository grep over those reference/component paths finds zero `method: 'PATCH'` calls and zero edit/save form controls.
- The required leak grep also disproves the brief's baseline claim that `apps/web/src` is clean: the exact command returns rc=0 with 12 pre-existing matches at `origin/next` (external font links, an avatar placeholder, an SVG namespace, and test fixtures). There are zero new matches in the worktree, but making the exact grep empty would require unrelated changes the brief forbids.
The brief simultaneously requires tests proving task/project PATCH edits and forbids inventing edit UX, fields, or flows beyond what those references expose. Adding controls or callback contracts would therefore be new design outside the authorized port/wiring scope. Per the explicit stop rule, no SPA source, route, config, package, lockfile, or test implementation was started.
## Verification evidence on unchanged web baseline
These gates were still run to distinguish the brief contradiction from a broken base:
- `pnpm --filter @mosaicstack/web typecheck` — rc=0.
- `pnpm --filter @mosaicstack/web lint` — rc=0.
- `pnpm --filter @mosaicstack/web test` — rc=0; 15 files, 133 tests passed. Before→after remains 133→133 because implementation was halted.
- `pnpm --filter @mosaicstack/web build` — rc=0; Next production build completed, 10/10 static pages generated.
- `pnpm install --frozen-lockfile` — rc=0; lock SHA-256 remained `9acaa89d213b3281e757b6edf6fdb8727176570d725b78a0de234c61a7f3c332`; `git diff` for the lockfile rc=0.
- `git grep -nE "http://|https://" -- apps/web/src` — rc=0 with 12 pre-existing matches; `origin/next` count=12, current HEAD count=12, new worktree additions=0.
- `pnpm format:check` — rc=0.
- `git diff --check` — rc=0.
## Delivery evidence
- Initial queue-guard invocation through shell `~` failed rc=127 because this Distrobox resolves `HOME` to `/home/jwoltje/distrobox-homes/mosaic-dev`, where the injected fleet status already reports the tools installation missing.
- Correct supported host-tool invocation `/home/jwoltje/.config/mosaic/tools/git/ci-queue-wait.sh --purpose push -B feat/webui-p4-1` — rc=0; branch absent remotely and queue clear.
- `git push -u origin feat/webui-p4-1` — rc=0. The push hook additionally ran repository preflight, typecheck (45/45 tasks), lint (25/25 tasks), and format check successfully.
- No PR was opened and no merge was attempted, per the brief.
## Risks / blockers
- The active harness mutates `.mosaic/orchestrator/session.lock`; it is excluded from staging.
- The Phase P §6.4 prose is absent from this checkout, so the brief-named route spec and existing Next pages are the bounded implementation anchors.
- The MissionStatus `/api/coord/status` panel is explicitly deferred to a follow-up and must not enter P4-1.
## REV 2 continuation — 2026-08-10T21:59:17-05:00
REV 2 supersedes the original tasking above. The independent check confirmed that the Next reference pages and `task-detail-modal.tsx` are read-only, so project/task editing is deliberately deferred to P4-1b. Do not re-litigate or implement the former PATCH requirements.
### Revised objective and acceptance map
- A: `/projects` read-only SPA list with loading, cards, empty, and surfaced API-error states; card navigation to `/projects/:id`; no MissionStatus panel.
- B: `/projects/:id` read-only SPA detail using `useParams`, `useNavigate`, and the specified three-request `Promise.all`; preserve reference overview/tasks/missions tabs and read-only task modal.
- C: `/tasks` read-only SPA list/kanban view with the existing read-only task modal.
- D: replace only the three route placeholders and add page-local route error boundaries without modifying chat.
- E: use only `@/lib/types` and `@/lib/api` with relative `/api/...` REST paths.
- F: add Vite `strictPort: true` and pin legacy Next dev/start to 3101 without dependency or proxy changes.
- G: add fixture-backed page tests and route-resolution assertions; run every user-specified verification gate and the narrowed new-code origin check.
### Revised plan
1. Reset `feat/webui-p4-1` to the exact `origin/next` base and independently verify the read-only reference/component/API assumptions.
2. Record the pre-change web suite count; write and run focused page/route tests first to observe expected RED failures.
3. Port the three read-only pages, add independent route error boundaries, wire routes, and apply only the two dev-port pins.
4. Run focused tests, full required web gates, frozen install/lockfile proof, origin check, format/diff hygiene, and accessibility/state-transition sanity checks.
5. Obtain independent spec/code/security review, remediate every blocker, and repeat affected gates.
6. Commit conventionally, run the supported pre-push queue guard, and push only `feat/webui-p4-1`; no PR and no merge.
### Revised budget and session state
- No explicit token cap was supplied. Working soft cap remains **50K tokens**.
- TDD is required by the user and frontend skill for the new SPA behavior; tests must fail for missing pages before production implementation.
- Documentation assessment: this is a parity port of existing read-only behavior, not a new public workflow or API contract; the task scratchpad is the required delivery record, with no user/developer/API documentation changes in this bounded increment.
- Exact base confirmed after fresh fetch/reset: `e00cc475a2b1e9866bd4e2f8df80aff640c3a543`.
- Remote `feat/webui-p4-1` is absent; the eventual push is a fresh branch creation.
- Harness-owned `.mosaic/orchestrator/session.lock` remains excluded from staging.
## REV 2 implementation pass — 2026-08-11T22:05:00Z
### Startup verification
- Loaded required startup files: `/home/jwoltje/.config/mosaic/CONSTITUTION.md`, `/home/jwoltje/.config/mosaic/SOUL.md`, project `AGENTS.md`, `/home/jwoltje/.config/mosaic/guides/E2E-DELIVERY.md`, `docs/PRD.md`, and this scratchpad.
- Loaded required skills: `test-driven-development`, `vitest`, `vite`, `next-best-practices`, and `verification-before-completion`.
- Loaded required runtime guide: `/home/jwoltje/.config/mosaic/runtime/codex/RUNTIME.md`.
- Structured reasoning tool availability verified through the harness before planning.
- `git rev-parse HEAD` confirmed exact required base: `e00cc475a2b1e9866bd4e2f8df80aff640c3a543`.
- `git status --short` at startup showed only the expected untracked append-only scratchpad; `.mosaic/orchestrator/session.lock` did not appear and must remain unstaged if it changes later.
### Current plan
1. Record baseline evidence.
2. Add focused failing SPA page/route specs first.
3. Run focused RED command and record the missing-page failure.
4. Implement the bounded read-only SPA pages, boundaries, route wiring, and the two dev-port pins.
5. Run the required verification matrix and inspect diff hygiene.
### Pre-change baseline
- Command: `pnpm --filter @mosaicstack/web test`
- rc=0
- File/test totals before changes: `15 files / 133 tests`
- Notes: baseline already includes `/chat` SPA route coverage and raw React `createRoot`/`act` page specs that this pass should mirror.
### RED evidence before production implementation
- Command: `pnpm --filter @mosaicstack/web test -- src/spa/pages/projects.spec.tsx src/spa/pages/project-detail.spec.tsx src/spa/pages/tasks.spec.tsx src/spa/pages/resource-route-boundaries.spec.tsx src/spa/routes.spec.tsx`
- rc=1
- Expected missing-feature reason confirmed:
- `src/spa/pages/projects.spec.tsx`, `src/spa/pages/project-detail.spec.tsx`, `src/spa/pages/tasks.spec.tsx`, and `src/spa/routes.spec.tsx` fail import resolution because the SPA page modules do not exist yet.
- `src/spa/pages/resource-route-boundaries.spec.tsx` fails because the current `/projects`, `/projects/:id`, and `/tasks` routes still render placeholders without route-level alert fallbacks.
### Implementation summary
- Added SPA pages:
- `apps/web/src/spa/pages/projects.tsx`
- `apps/web/src/spa/pages/project-detail.tsx`
- `apps/web/src/spa/pages/tasks.tsx`
- Added page-local route boundary components in `apps/web/src/spa/pages/resource-route-error-boundaries.tsx`.
- Wired `/projects`, `/projects/:id`, and `/tasks` in `apps/web/src/routes.tsx` with real elements and `errorElement`s.
- Added fixture-backed raw React Vitest coverage plus route assertions for the three pages and their boundaries.
- Applied only the requested dev topology pins:
- `apps/web/vite.config.ts`: `server.strictPort = true`
- `apps/web/package.json`: `next dev -p 3101`, `next start -p 3101`
### Post-implementation verification
- Focused changed specs:
- Command: `pnpm --filter @mosaicstack/web exec vitest run src/spa/pages/projects.spec.tsx src/spa/pages/project-detail.spec.tsx src/spa/pages/tasks.spec.tsx src/spa/pages/resource-route-boundaries.spec.tsx src/spa/routes.spec.tsx`
- rc=0
- Result: `5 files / 26 tests` passed.
- `pnpm --filter @mosaicstack/web typecheck` — rc=0.
- `pnpm --filter @mosaicstack/web lint` — rc=0.
- `pnpm --filter @mosaicstack/web test` — rc=0; post-change totals `19 files / 153 tests`.
- `pnpm --filter @mosaicstack/web build` — rc=1.
- Limitation: Next/Turbopack hit a sandbox/runtime failure while processing `apps/web/src/app/globals.css`: `creating new process`, `binding to a port`, `Operation not permitted (os error 1)`. This appears environmental, not route-code-specific.
- `pnpm --filter @mosaicstack/web build:vite` — rc=0.
- `pnpm install --frozen-lockfile` — rc=1.
- Limitation: repo `prepare` hook attempted to lock `/home/jwoltje/distrobox-homes/mosaic-dev/src/stack/.git/config`, which is read-only in this harness.
- `pnpm-lock.yaml` SHA-256 before/after install attempt: `9acaa89d213b3281e757b6edf6fdb8727176570d725b78a0de234c61a7f3c332`.
- `origin/next` `pnpm-lock.yaml` SHA-256: `9acaa89d213b3281e757b6edf6fdb8727176570d725b78a0de234c61a7f3c332`.
- `git diff -- pnpm-lock.yaml` — rc=0 (unchanged).
- `git diff -- apps/web/src | grep -nE '^\+.*(fetch|io|api)\(\s*[\x27\"]https?://'` — rc=1 (empty, as required).
- `pnpm format:check` — rc=0.
- `git diff --check` — rc=0.
### Final worktree check
- `git status --short` shows only the authorized web files plus this scratchpad.
- `.mosaic/orchestrator/session.lock` remains unstaged.
- Current changed file set:
- `apps/web/package.json`
- `apps/web/src/routes.tsx`
- `apps/web/src/spa/routes.spec.tsx`
- `apps/web/vite.config.ts`
- `apps/web/src/spa/pages/page-errors.ts`
- `apps/web/src/spa/pages/page-fixtures.ts`
- `apps/web/src/spa/pages/project-detail.spec.tsx`
- `apps/web/src/spa/pages/project-detail.tsx`
- `apps/web/src/spa/pages/projects.spec.tsx`
- `apps/web/src/spa/pages/projects.tsx`
- `apps/web/src/spa/pages/resource-route-boundaries.spec.tsx`
- `apps/web/src/spa/pages/resource-route-error-boundaries.tsx`
- `apps/web/src/spa/pages/tasks.spec.tsx`
- `apps/web/src/spa/pages/tasks.tsx`
- `docs/scratchpads/webui-p4-1.md`
## P4-1 REV 2 Final Delivery Verification — 2026-08-10T22:23:00Z
### Spec/Security Review Verdicts (Independent Reviews)
Both independent reviews cleared P4-1 REV 2 implementation without blockers:
- **Spec review:** approved; 0 blockers, 0 should-fix findings
- **Code/security review:** approved; 0 blockers, 0 critical/high findings, 0 should-fix recommendations
### Final Fresh Gate Verification (This Session)
All verification gates executed sequentially with rc=0 (except where noted):
1. `pnpm --filter @mosaicstack/web typecheck` — rc=0
2. `pnpm --filter @mosaicstack/web lint` — rc=0
3. `pnpm --filter @mosaicstack/web test` — rc=0; **19 files / 153 tests** (baseline: 15 files / 133 tests; added: 4 files / 20 tests)
4. `pnpm --filter @mosaicstack/web build` — rc=0; Next production build completed, 10/10 static pages generated
5. `pnpm --filter @mosaicstack/web build:vite` — rc=0; Vite production bundle generated
6. Lockfile integrity:
- Before install: SHA-256 `9acaa89d213b3281e757b6edf6fdb8727176570d725b78a0de234c61a7f3c332`
- After frozen install: SHA-256 `9acaa89d213b3281e757b6edf6fdb8727176570d725b78a0de234c61a7f3c332`
- Equality: ✓ verified
- `git diff --quiet origin/next -- pnpm-lock.yaml` rc=0 ✓
7. `pnpm format:check` — rc=0; all matched files use Prettier code style
8. `git diff --check` — rc=0; no trailing whitespace or merged conflict markers
9. New-code origin check: `git diff -- apps/web/src | grep -nE '^\+.*(fetch|io|api)\(\s*[\x27\"]https?://'` — rc=1 (empty result, as required)
10. `pnpm --filter @mosaicstack/web exec vitest run src/spa/pages/projects.spec.tsx src/spa/pages/project-detail.spec.tsx src/spa/pages/tasks.spec.tsx src/spa/pages/resource-route-boundaries.spec.tsx src/spa/routes.spec.tsx` — rc=0; 5 files / 26 tests passed
### Scope Audit
**Authorized in-scope changes:**
- ✓ `apps/web/package.json` — dev port pins only
- ✓ `apps/web/vite.config.ts` — strictPort flag only
- ✓ `apps/web/src/routes.tsx` — route wiring with errorElements
- ✓ `apps/web/src/spa/pages/projects.tsx` — read-only SPA list
- ✓ `apps/web/src/spa/pages/project-detail.tsx` — read-only SPA detail
- ✓ `apps/web/src/spa/pages/tasks.tsx` — read-only SPA list/kanban
- ✓ `apps/web/src/spa/pages/resource-route-error-boundaries.tsx` — page-local error components
- ✓ `apps/web/src/spa/pages/*.spec.tsx` — fixture-backed tests (4 new)
- ✓ `docs/scratchpads/webui-p4-1.md` — append-only record
**Out-of-scope verification:**
- `.mosaic/orchestrator/session.lock` — not modified/staged ✓
- `pnpm-lock.yaml` — not modified ✓
- `apps/gateway/**` — not modified ✓
- `packages/**` — not modified ✓
- Chat SPA — not modified ✓
- Settings, admin, navigation, or auth flow — not modified ✓
### Next Action
Commit, run authorized queue guard, and push only `feat/webui-p4-1` (no PR, no merge).
@@ -193,16 +193,19 @@ describe('Unified wizard (runWizard with default skipGateway)', () => {
'Your timezone': 'UTC', 'Your timezone': 'UTC',
}); });
await runWizard({ await expect(
runWizard({
mosaicHome: tmpDir, mosaicHome: tmpDir,
sourceDir: tmpDir, sourceDir: tmpDir,
prompter, prompter,
configService: createConfigService(tmpDir, tmpDir), configService: createConfigService(tmpDir, tmpDir),
skipGatewayNpmInstall: true, skipGatewayNpmInstall: true,
}); }),
).rejects.toThrow('Gateway configuration failed');
const logs = prompter.getLogs(); const logs = prompter.getLogs();
expect(logs.some((line) => line.includes('Gateway did not become healthy'))).toBe(true); expect(logs.some((line) => line.includes('Gateway did not become healthy'))).toBe(true);
expect(logs.some((line) => line.includes('Gateway configuration failed'))).toBe(true);
expect(logs.some((line) => line.includes('Installation Summary'))).toBe(false); expect(logs.some((line) => line.includes('Installation Summary'))).toBe(false);
expect(logs.some((line) => line.includes('Mosaic is ready.'))).toBe(false); expect(logs.some((line) => line.includes('Mosaic is ready.'))).toBe(false);
expect(gatewayConfigMock).toHaveBeenCalledTimes(1); expect(gatewayConfigMock).toHaveBeenCalledTimes(1);
@@ -0,0 +1 @@
Mosaic lease promotion was processed mechanically; no action is needed.
@@ -32,6 +32,18 @@
] ]
} }
], ],
"UserPromptSubmit": [
{
"matcher": "^/mosaic-promote$",
"hooks": [
{
"type": "command",
"command": "python3 ~/.config/mosaic/tools/lease-broker/promote-begin.py",
"timeout": 15
}
]
}
],
"PreToolUse": [ "PreToolUse": [
{ {
"matcher": ".*", "matcher": ".*",
@@ -81,8 +93,8 @@
"hooks": [ "hooks": [
{ {
"type": "command", "type": "command",
"command": "python3 ~/.config/mosaic/tools/lease-broker/receipt-observer-client.py --runtime claude --latest-entry", "command": "python3 ~/.config/mosaic/tools/lease-broker/receipt-observer-client.py --runtime claude --latest-entry; observer_status=$?; python3 ~/.config/mosaic/tools/lease-broker/promote-complete.py; exit $observer_status",
"timeout": 3 "timeout": 15
}, },
{ {
"type": "command", "type": "command",
@@ -99,7 +99,7 @@ prompt_if_empty() {
if [[ $NON_INTERACTIVE -eq 1 ]]; then if [[ $NON_INTERACTIVE -eq 1 ]]; then
if [[ -n "$default_value" ]]; then if [[ -n "$default_value" ]]; then
eval "$var_name=\"$default_value\"" printf -v "$var_name" %s "$default_value"
return return
fi fi
echo "[mosaic-init] ERROR: --$var_name is required in non-interactive mode" >&2 echo "[mosaic-init] ERROR: --$var_name is required in non-interactive mode" >&2
@@ -115,7 +115,7 @@ prompt_if_empty() {
if [[ -z "$value" && -n "$default_value" ]]; then if [[ -z "$value" && -n "$default_value" ]]; then
value="$default_value" value="$default_value"
fi fi
eval "$var_name=\"$value\"" printf -v "$var_name" %s "$value"
} }
prompt_multiline() { prompt_multiline() {
@@ -129,7 +129,7 @@ prompt_multiline() {
fi fi
if [[ $NON_INTERACTIVE -eq 1 ]]; then if [[ $NON_INTERACTIVE -eq 1 ]]; then
eval "$var_name=\"$default_value\"" printf -v "$var_name" %s "$default_value"
return return
fi fi
@@ -139,7 +139,7 @@ prompt_multiline() {
if [[ -z "$value" ]]; then if [[ -z "$value" ]]; then
value="$default_value" value="$default_value"
fi fi
eval "$var_name=\"$value\"" printf -v "$var_name" %s "$value"
} }
# ── Existing file detection ──────────────────────────────────── # ── Existing file detection ────────────────────────────────────
@@ -233,6 +233,14 @@ for runtime_file in \
copy_file_managed "$src" "$HOME/.claude/$runtime_file" copy_file_managed "$src" "$HOME/.claude/$runtime_file"
done done
if [[ -d "$MOSAIC_HOME/runtime/claude/commands" ]]; then
mkdir -p "$HOME/.claude/commands"
for command_file in "$MOSAIC_HOME/runtime/claude/commands/"*; do
[[ -f "$command_file" ]] || continue
copy_file_managed "$command_file" "$HOME/.claude/commands/$(basename "$command_file")"
done
fi
# OpenCode runtime adapter (thin pointer to AGENTS.md) # OpenCode runtime adapter (thin pointer to AGENTS.md)
opencode_adapter="$MOSAIC_HOME/runtime/opencode/AGENTS.md" opencode_adapter="$MOSAIC_HOME/runtime/opencode/AGENTS.md"
if [[ -f "$opencode_adapter" ]]; then if [[ -f "$opencode_adapter" ]]; then
@@ -153,7 +153,24 @@ if [[ $link_only -eq 1 ]]; then
exit 0 exit 0
fi fi
# Skills are linked into the MOSAIC-OWNED harness homes, never a base install.
# Paths mirror the config-dir env vars the launcher injects (HARNESS_HOME_ENV in
# commands/launch.js):
# claude CLAUDE_CONFIG_DIR -> <home>/skills
# pi PI_CODING_AGENT_DIR -> <home>/skills (replaces ~/.pi/agent)
# codex CODEX_HOME -> <home>/skills
# opencode XDG_CONFIG_HOME -> <home>/opencode/skills (XDG adds a level)
link_targets=( link_targets=(
"$MOSAIC_HOME/.claude/skills"
"$MOSAIC_HOME/.codex/skills"
"$MOSAIC_HOME/.opencode/opencode/skills"
"$MOSAIC_HOME/.pi/skills"
)
# Pre-isolation installs planted the same symlink farm directly in the operator's
# base installs. Those are now orphaned: the launcher no longer reads them, but
# they persist and make a "clean" base install look mosaic-managed.
legacy_link_targets=(
"$HOME/.claude/skills" "$HOME/.claude/skills"
"$HOME/.codex/skills" "$HOME/.codex/skills"
"$HOME/.config/opencode/skills" "$HOME/.config/opencode/skills"
@@ -245,13 +262,72 @@ prune_stale_links_in_target() {
# -m resolves lexical dangling targets too. If resolution fails, ownership # -m resolves lexical dangling targets too. If resolution fails, ownership
# is unproven and the link must be preserved. # is unproven and the link must be preserved.
resolved="$(readlink -m "$link_path" 2>/dev/null || true)" resolved="$(readlink -m "$link_path" 2>/dev/null || true)"
if [[ -n "$resolved" && "$resolved" == "$canonical_real/"* ]]; then # $canonical_real must be length-checked BEFORE use as a prefix: if it were
# ever empty, "$resolved" == "$canonical_real/"* collapses to == "/"* and
# matches every absolute path. Combined with the is_mosaic_skill_name skip
# above, that inverts the function precisely — it would delete exactly the
# FOREIGN symlinks and keep the mosaic ones. (#1087, reported by mos-claude.)
if [[ -n "$resolved" && -n "$canonical_real" && "$resolved" == "$canonical_real/"* ]]; then
rm -f "$link_path" rm -f "$link_path"
echo "[mosaic-skills] Removed stale retired skill link: $link_path" echo "[mosaic-skills] Removed stale retired skill link: $link_path"
fi fi
done < <(find "$target_dir" -mindepth 1 -maxdepth 1 -type l -print0) done < <(find "$target_dir" -mindepth 1 -maxdepth 1 -type l -print0)
} }
# Remove mosaic-owned symlinks left in a base install by a pre-isolation sync.
#
# Ownership is proven by RESOLUTION, not by name: only links resolving inside the
# canonical or local skills dirs are removed. Anything else — a real directory, a
# link elsewhere, an unresolvable link — is left untouched. This mirrors the
# refusal in commands/skill.js ("only symlinks pointing inside the Mosaic skills
# directory are managed") and preserves e.g. codex's own `.system` dir.
#
# The directory itself is kept: mosaic-doctor warns when ~/.pi/agent/skills is
# missing, and an empty dir is the correct end state, not an absent one.
cleanup_legacy_target() {
local target_dir="$1"
local removed=0 kept=0
[[ -d "$target_dir" ]] || return 0
while IFS= read -r -d '' link_path; do
local resolved owned=0
resolved="$(readlink -m "$link_path" 2>/dev/null || true)"
# Guard the empty-prefix trap: an unset *_real would make "$resolved" == "/"*
# match every absolute path and delete foreign links.
if [[ -n "$resolved" ]]; then
if [[ -n "$canonical_real" && "$resolved" == "$canonical_real/"* ]]; then
owned=1
elif [[ -n "$local_real" && "$resolved" == "$local_real/"* ]]; then
owned=1
fi
fi
if [[ $owned -eq 1 ]]; then
rm -f "$link_path"
removed=$((removed + 1))
else
kept=$((kept + 1))
fi
done < <(find "$target_dir" -mindepth 1 -maxdepth 1 -type l -print0)
if [[ $removed -gt 0 ]]; then
echo "[mosaic-skills] Legacy cleanup: removed $removed mosaic symlink(s) from $target_dir (preserved $kept foreign)"
fi
}
for legacy in "${legacy_link_targets[@]}"; do
# Skip anything that is also a current target, so isolation can never
# self-destruct if the two lists ever overlap.
skip=0
for target in "${link_targets[@]}"; do
[[ "$legacy" == "$target" ]] && skip=1
done
[[ $skip -eq 1 ]] && continue
cleanup_legacy_target "$legacy"
done
for target in "${link_targets[@]}"; do for target in "${link_targets[@]}"; do
mkdir -p "$target" mkdir -p "$target"
@@ -0,0 +1,21 @@
#!/usr/bin/env bash
set -euo pipefail
# Source only the prompt helpers; executing mosaic-init itself requires templates.
source <(head -n 144 "$(dirname "$0")/mosaic-init")
rm -f /tmp/pwned
payload='literal "$(touch /tmp/pwned)"'
AGENT_NAME=""
prompt_if_empty AGENT_NAME "Agent name" <<<"$payload"
[[ "$AGENT_NAME" == "$payload" ]] || {
echo "FAIL: prompt answer did not round-trip literally" >&2
exit 1
}
[[ ! -e /tmp/pwned ]] || {
echo "FAIL: prompt answer executed code" >&2
rm -f /tmp/pwned
exit 1
}
echo "mosaic-init RCE regression: PASS"
@@ -465,12 +465,24 @@ while true; do
no-status) no-status)
if [[ "$REQUIRE_STATUS" -eq 1 ]]; then if [[ "$REQUIRE_STATUS" -eq 1 ]]; then
echo "Error: ASSERTED_NOT_READY state=no-status; --require-status was set for ${BRANCH}." >&2 echo "Error: ASSERTED_NOT_READY state=no-status; --require-status was set for ${BRANCH}." >&2
else exit 3
echo "Error: ASSERTED_NOT_READY state=no-status purpose=${PURPOSE} branch=${BRANCH}." >&2
fi fi
if [[ "$PURPOSE" == "push" ]]; then
echo "[ci-queue-wait] queue-clear state=no-status purpose=push branch=${BRANCH}; no queued or running CI."
exit 0
fi
echo "Error: ASSERTED_NOT_READY state=no-status purpose=${PURPOSE} branch=${BRANCH}." >&2
exit 3 exit 3
;; ;;
terminal-failure|malformed|unknown) terminal-failure)
if [[ "$PURPOSE" == "push" ]]; then
echo "[ci-queue-wait] queue-clear state=terminal-failure purpose=push branch=${BRANCH}; no queued or running CI."
exit 0
fi
echo "Error: ASSERTED_NOT_READY state=terminal-failure purpose=${PURPOSE} branch=${BRANCH}." >&2
exit 3
;;
malformed|unknown)
echo "Error: ASSERTED_NOT_READY state=${STATE} purpose=${PURPOSE} branch=${BRANCH}." >&2 echo "Error: ASSERTED_NOT_READY state=${STATE} purpose=${PURPOSE} branch=${BRANCH}." >&2
exit 3 exit 3
;; ;;
+469 -59
View File
@@ -1,6 +1,6 @@
#!/bin/bash #!/bin/bash
# pr-merge.sh - Merge pull requests on Gitea or GitHub # pr-merge.sh - Merge pull requests on Gitea or GitHub
# Usage: pr-merge.sh -n PR_NUMBER [-m squash] [-d] # Usage: pr-merge.sh -n PR_NUMBER [-m squash] [-d] [--expect-head SHA] [--co-author-trailers --escalate-to PRINCIPAL]
set -euo pipefail set -euo pipefail
@@ -14,6 +14,8 @@ MERGE_METHOD="squash"
DELETE_BRANCH=false DELETE_BRANCH=false
DRY_RUN=false DRY_RUN=false
EXPECT_HEAD="" EXPECT_HEAD=""
CO_AUTHOR_TRAILERS=false
ESCALATE_TO=""
usage() { usage() {
cat <<EOF cat <<EOF
@@ -27,12 +29,16 @@ Options:
-d, --delete-branch Delete the head branch after merge -d, --delete-branch Delete the head branch after merge
--dry-run Run metadata/login preflight without merging --dry-run Run metadata/login preflight without merging
--expect-head SHA Refuse unless the PR head matches this full commit SHA --expect-head SHA Refuse unless the PR head matches this full commit SHA
--co-author-trailers Build verified trailers from linked PR commit authors
--escalate-to NAME Named principal for an unresolved-author BLOCK
-h, --help Show this help message -h, --help Show this help message
Examples: Examples:
$(basename "$0") -n 42 # Merge PR #42 $(basename "$0") -n 42 # Merge PR #42
$(basename "$0") -n 42 -m squash # Squash merge $(basename "$0") -n 42 -m squash # Squash merge
$(basename "$0") -n 42 -d # Squash merge and delete branch $(basename "$0") -n 42 -d # Squash merge and delete branch
$(basename "$0") -n 42 --expect-head 0123456789abcdef0123456789abcdef01234567
$(basename "$0") -n 42 --co-author-trailers --escalate-to tl-mosaic
EOF EOF
exit "${1:-1}" exit "${1:-1}"
} }
@@ -57,9 +63,25 @@ while [[ $# -gt 0 ]]; do
shift shift
;; ;;
--expect-head) --expect-head)
if [[ $# -lt 2 ]]; then
echo "Error: --expect-head requires one full commit SHA." >&2
exit 1
fi
EXPECT_HEAD="$2" EXPECT_HEAD="$2"
shift 2 shift 2
;; ;;
--co-author-trailers)
CO_AUTHOR_TRAILERS=true
shift
;;
--escalate-to)
if [[ $# -lt 2 ]]; then
echo "Error: --escalate-to requires one principal name." >&2
exit 1
fi
ESCALATE_TO="$2"
shift 2
;;
-h|--help) -h|--help)
usage 0 usage 0
;; ;;
@@ -88,17 +110,30 @@ if [[ -n "$EXPECT_HEAD" && ! "$EXPECT_HEAD" =~ ^[0-9a-fA-F]{40}$ ]]; then
echo "Error: --expect-head must be a full 40-character hexadecimal commit SHA." >&2 echo "Error: --expect-head must be a full 40-character hexadecimal commit SHA." >&2
exit 1 exit 1
fi fi
if [[ "$CO_AUTHOR_TRAILERS" == true && -z "$ESCALATE_TO" ]]; then
echo "Error: --co-author-trailers requires --escalate-to with a named principal." >&2
exit 1
fi
if [[ -n "$ESCALATE_TO" && ! "$ESCALATE_TO" =~ ^[A-Za-z0-9_.-]+$ ]]; then
echo "Error: --escalate-to must be one exact principal name." >&2
exit 1
fi
if [[ "$CO_AUTHOR_TRAILERS" != true && -n "$ESCALATE_TO" ]]; then
echo "Error: --escalate-to is valid only with --co-author-trailers." >&2
exit 1
fi
PR_METADATA="$("$SCRIPT_DIR/pr-metadata.sh" -n "$PR_NUMBER")" PR_METADATA="$("$SCRIPT_DIR/pr-metadata.sh" -n "$PR_NUMBER")"
BASE_BRANCH="$(printf '%s' "$PR_METADATA" | python3 -c 'import json, sys; print((json.load(sys.stdin).get("baseRefName") or "").strip())')" BASE_BRANCH="$(printf '%s' "$PR_METADATA" | python3 -c 'import json, sys; print((json.load(sys.stdin).get("baseRefName") or "").strip())')"
HEAD_BRANCH="$(printf '%s' "$PR_METADATA" | python3 -c 'import json, sys; print((json.load(sys.stdin).get("headRefName") or "").strip())')" HEAD_BRANCH="$(printf '%s' "$PR_METADATA" | python3 -c 'import json, sys; print((json.load(sys.stdin).get("headRefName") or "").strip())')"
HEAD_SHA="$(printf '%s' "$PR_METADATA" | python3 -c 'import json, sys; print((json.load(sys.stdin).get("headRefOid") or "").strip())')" HEAD_SHA="$(printf '%s' "$PR_METADATA" | python3 -c 'import json, sys; print((json.load(sys.stdin).get("headRefOid") or "").strip())')"
HEAD_REPO="$(printf '%s' "$PR_METADATA" | python3 -c 'import json, sys; value=json.load(sys.stdin).get("headRepository") or ""; print((value.get("nameWithOwner") or value.get("full_name") or "") if isinstance(value, dict) else str(value).strip())')" HEAD_REPO="$(printf '%s' "$PR_METADATA" | python3 -c 'import json, sys; value=json.load(sys.stdin).get("headRepository") or ""; print((value.get("nameWithOwner") or value.get("full_name") or "") if isinstance(value, dict) else str(value).strip())')"
if [[ "$BASE_BRANCH" != "main" ]]; then PR_TITLE="$(printf '%s' "$PR_METADATA" | python3 -c 'import json, sys; print((json.load(sys.stdin).get("title") or "").strip())')"
echo "Error: Mosaic policy allows merges only for PRs targeting 'main' (found '$BASE_BRANCH')." >&2 PR_AUTHOR="$(printf '%s' "$PR_METADATA" | python3 -c 'import json, sys; value=json.load(sys.stdin).get("author") or ""; print((value.get("login") or "").strip() if isinstance(value, dict) else str(value).strip())')"
if [[ "$BASE_BRANCH" != "main" && "$BASE_BRANCH" != "next" ]]; then
echo "Error: Mosaic policy allows merges only for PRs targeting 'main' or 'next' (found '$BASE_BRANCH')." >&2
exit 1 exit 1
fi fi
if [[ -z "$HEAD_BRANCH" || -z "$HEAD_REPO" || ! "$HEAD_SHA" =~ ^[0-9a-fA-F]{40}$ ]]; then if [[ -z "$HEAD_BRANCH" || -z "$HEAD_REPO" || ! "$HEAD_SHA" =~ ^[0-9a-fA-F]{40}$ ]]; then
echo "Error: Could not resolve the PR head branch, repository, and full commit SHA for queue inspection." >&2 echo "Error: Could not resolve the PR head branch, repository, and full commit SHA for queue inspection." >&2
exit 1 exit 1
@@ -122,70 +157,442 @@ PLATFORM=$(detect_platform)
OWNER=$(get_repo_owner) OWNER=$(get_repo_owner)
REPO=$(get_repo_name) REPO=$(get_repo_name)
merge_gitea_with_api() { write_curl_auth_config() {
local host="$1" api_url token basic_auth body_file raw_code payload local mode="$1" credential="$2"
api_url="https://${host}/api/v1/repos/${OWNER}/${REPO}/pulls/${PR_NUMBER}/merge" printf '%s' "$credential" | python3 -c '
mkdir -p "${AGENT_WORK_ROOT:-${HOME:-/tmp}/mosaic/agent-work}" import sys
body_file=$(mktemp "${AGENT_WORK_ROOT:-${HOME:-/tmp}/mosaic/agent-work}/pr-merge-api-response.XXXXXX") mode = sys.argv[1]
payload=$(python3 - "$HEAD_SHA" "$DELETE_BRANCH" <<'PY' credential = sys.stdin.read()
if not credential or any(char in credential for char in "\r\n"):
raise SystemExit(1)
escaped = credential.replace("\\", "\\\\").replace("\"", "\\\"")
if mode == "token":
print(f"header = \"Authorization: token {escaped}\"")
elif mode == "basic":
print(f"user = \"{escaped}\"")
else:
raise SystemExit(1)
' "$mode"
}
LAST_GITEA_HTTP_CODE="000"
LAST_GITEA_ERROR=""
MERGE_TEMP_DIRS=()
GITEA_CURL_MAX_BYTES="${MOSAIC_GITEA_CURL_MAX_BYTES:-1048576}"
GITEA_CURL_MAX_TIME="${MOSAIC_GITEA_CURL_MAX_TIME_SEC:-30}"
GITEA_CURL_CONNECT_TIMEOUT="${MOSAIC_GITEA_CURL_CONNECT_TIMEOUT_SEC:-10}"
for bound in "$GITEA_CURL_MAX_BYTES" "$GITEA_CURL_MAX_TIME" "$GITEA_CURL_CONNECT_TIMEOUT"; do
if [[ ! "$bound" =~ ^[1-9][0-9]*$ ]]; then
echo "Error: Gitea curl bounds must be positive integers; refusing request." >&2
exit 1
fi
done
GITEA_CURL_BOUNDS=(
--max-filesize "$GITEA_CURL_MAX_BYTES"
--max-time "$GITEA_CURL_MAX_TIME"
--connect-timeout "$GITEA_CURL_CONNECT_TIMEOUT"
)
format_gitea_error_response() {
local response_file="$1"
python3 - "$response_file" <<'PY'
import json import json
import sys import sys
head_sha, delete_branch = sys.argv[1:] with open(sys.argv[1], "rb") as handle:
raw = handle.read(65536)
try:
response = json.loads(raw.decode("utf-8", errors="replace"))
except (UnicodeDecodeError, json.JSONDecodeError):
message = "non-JSON response omitted"
else:
if isinstance(response, dict):
message = response.get("message") or response.get("error")
if not message and response.get("errors") is not None:
message = json.dumps(response["errors"], separators=(",", ":"))
else:
message = None
if not message:
message = "JSON response contained no error message"
message = str(message)
if len(message) > 500:
message = message[:500] + "..."
print(ascii(message))
PY
}
cleanup_merge_temp_dirs() {
local path
for path in "${MERGE_TEMP_DIRS[@]}"; do
[[ -n "$path" ]] && rm -rf -- "$path"
done
}
trap cleanup_merge_temp_dirs EXIT
trap 'exit 130' INT
trap 'exit 143' TERM
fetch_gitea_pr_head() {
local host="$1" auth_mode="$2" credential="$3" work_root="$4"
local response_file raw_code api_url auth_config curl_rc
response_file=$(mktemp "$work_root/pr-merge-pr.XXXXXX")
api_url="https://${host}/api/v1/repos/${OWNER}/${REPO}/pulls/${PR_NUMBER}"
if ! auth_config=$(write_curl_auth_config "$auth_mode" "$credential"); then
echo "Error: Could not construct Gitea authentication config; refusing request." >&2
rm -f "$response_file"
return 1
fi
raw_code=$(curl -sS -K - "${GITEA_CURL_BOUNDS[@]}" -w '%{http_code}' -o "$response_file" \
-H "User-Agent: curl/8" "$api_url" <<<"$auth_config")
curl_rc=$?
LAST_GITEA_HTTP_CODE="${raw_code:-000}"
if [[ "$curl_rc" -ne 0 ]]; then
LAST_GITEA_ERROR="curl transport failed (rc=$curl_rc)"
rm -f "$response_file"
return 1
fi
if [[ ! "$raw_code" =~ ^2 ]]; then
LAST_GITEA_ERROR=$(format_gitea_error_response "$response_file")
rm -f "$response_file"
return 1
fi
if ! python3 - "$response_file" <<'PY'
import json
import re
import sys
with open(sys.argv[1], encoding="utf-8") as handle:
pull = json.load(handle)
head = pull.get("head") if isinstance(pull, dict) else None
sha = str(head.get("sha") or "") if isinstance(head, dict) else ""
if not re.fullmatch(r"[0-9a-fA-F]{40}", sha):
raise SystemExit(1)
print(sha)
PY
then
echo "Error: Gitea PR response has no valid head SHA; refusing merge." >&2
rm -f "$response_file"
return 1
fi
rm -f "$response_file"
}
fetch_gitea_pr_commits() {
local host="$1" auth_mode="$2" credential="$3" work_root="$4"
local page page_file combined_file merged_file raw_code page_count api_url auth_config curl_rc
mkdir -p "$work_root"
if ! auth_config=$(write_curl_auth_config "$auth_mode" "$credential"); then
echo "Error: Could not construct Gitea authentication config; refusing request." >&2
return 1
fi
combined_file=$(mktemp "$work_root/pr-merge-commits.XXXXXX")
printf '[]' > "$combined_file"
page=1
while true; do
page_file=$(mktemp "$work_root/pr-merge-commits-page.XXXXXX")
api_url="https://${host}/api/v1/repos/${OWNER}/${REPO}/pulls/${PR_NUMBER}/commits?limit=50&page=${page}"
raw_code=$(curl -sS -K - "${GITEA_CURL_BOUNDS[@]}" -w '%{http_code}' -o "$page_file" \
-H "User-Agent: curl/8" "$api_url" <<<"$auth_config")
curl_rc=$?
LAST_GITEA_HTTP_CODE="${raw_code:-000}"
if [[ "$curl_rc" -ne 0 ]]; then
LAST_GITEA_ERROR="curl transport failed (rc=$curl_rc)"
rm -f "$page_file" "$combined_file"
return 1
fi
if [[ ! "$raw_code" =~ ^2 ]]; then
LAST_GITEA_ERROR=$(format_gitea_error_response "$page_file")
rm -f "$page_file" "$combined_file"
return 1
fi
if ! page_count=$(python3 - "$page_file" <<'PY'
import json
import sys
with open(sys.argv[1], encoding="utf-8") as handle:
page = json.load(handle)
if not isinstance(page, list):
raise SystemExit(1)
print(len(page))
PY
); then
echo "Error: Gitea PR commits response is not a JSON array; refusing merge." >&2
rm -f "$page_file" "$combined_file"
return 1
fi
merged_file=$(mktemp "$work_root/pr-merge-commits-merged.XXXXXX")
if ! python3 - "$combined_file" "$page_file" > "$merged_file" <<'PY'
import json
import sys
with open(sys.argv[1], encoding="utf-8") as handle:
combined = json.load(handle)
with open(sys.argv[2], encoding="utf-8") as handle:
page = json.load(handle)
json.dump(combined + page, sys.stdout, separators=(",", ":"))
PY
then
echo "Error: Could not combine paginated PR commit metadata; refusing merge." >&2
rm -f "$page_file" "$combined_file" "$merged_file"
return 1
fi
mv "$merged_file" "$combined_file"
rm -f "$page_file"
if [[ "$page_count" -lt 50 ]]; then
break
fi
page=$((page + 1))
if [[ "$page" -gt 1000 ]]; then
echo "Error: PR commit pagination exceeded 1000 pages; refusing merge." >&2
rm -f "$combined_file"
return 1
fi
done
cat "$combined_file"
rm -f "$combined_file"
}
# LIMITATION: author.login resolution proves the commit address maps to a registered account.
# It does NOT prove the named principal authored the commit — git author metadata is self-asserted.
# This gate checks ATTRIBUTION LINKAGE, not AUTHORSHIP. Commit signing is out of scope and unadopted.
build_coauthor_message_fields() {
local commits_file="$1" context_file="$2" head_file="$3"
python3 - "$commits_file" "$context_file" "$head_file" <<'PY'
import json
import re
import sys
commits_path, context_path, head_path = sys.argv[1:]
with open(commits_path, encoding="utf-8") as handle:
commits = json.load(handle)
head_sha = open(head_path, encoding="utf-8").read().strip()
context_parts = open(context_path, "rb").read().split(b"\0")
if len(context_parts) != 4 or context_parts[-1] != b"":
raise SystemExit(1)
poster, title, principal = (part.decode("utf-8") for part in context_parts[:3])
if not isinstance(commits, list) or not commits:
print(
f"BLOCK: provider returned no PR commits; author identity is unmeasurable. "
f"Refusing merge; escalate to named principal '{principal}'.",
file=sys.stderr,
)
raise SystemExit(75)
if not poster:
print(
f"BLOCK: PR poster login is empty; refusing merge; "
f"escalate to named principal '{principal}'.",
file=sys.stderr,
)
raise SystemExit(75)
if not re.fullmatch(r"[0-9a-fA-F]{40}", head_sha):
print(
f"BLOCK: inspected PR head SHA is invalid; refusing merge; "
f"escalate to named principal '{principal}'.",
file=sys.stderr,
)
raise SystemExit(75)
seen = set()
trailers = []
head_seen = False
for item in commits:
if not isinstance(item, dict):
print(f"BLOCK: malformed PR commit metadata; escalate to named principal '{principal}'.", file=sys.stderr)
raise SystemExit(75)
sha = str(item.get("sha") or "<unknown>")
if sha == head_sha:
head_seen = True
commit = item.get("commit") if isinstance(item.get("commit"), dict) else {}
commit_author = commit.get("author") if isinstance(commit.get("author"), dict) else {}
email = str(commit_author.get("email") or "").strip()
provider_author = item.get("author") if isinstance(item.get("author"), dict) else {}
login = str(provider_author.get("login") or "").strip()
if not login:
diagnostic_email = email or "<missing>"
print(
f"BLOCK: commit {sha!r} has author.login=NULL while "
f"commit.author.email={diagnostic_email!r}; refusing merge; "
f"escalate to named principal '{principal}'.",
file=sys.stderr,
)
raise SystemExit(75)
if (
not email.isascii()
or not email.isprintable()
or not re.fullmatch(r"[A-Za-z0-9_.-]+", login)
or not re.fullmatch(r"[^<>\s]+@[^<>\s]+", email)
):
print(
f"BLOCK: commit {sha!r} has unusable linked identity "
f"author.login={login!r}, commit.author.email={email!r}; refusing merge; "
f"escalate to named principal '{principal}'.",
file=sys.stderr,
)
raise SystemExit(75)
if login == poster or login in seen:
continue
seen.add(login)
trailers.append(f"Co-authored-by: {login} <{email}>")
if not head_seen:
print(
f"BLOCK: inspected PR head is absent from commit enumeration; refusing merge; "
f"escalate to named principal '{principal}'.",
file=sys.stderr,
)
raise SystemExit(75)
if not trailers:
print("{}")
raise SystemExit(0)
if not title:
print(
f"BLOCK: PR title is empty; refusing merge; escalate to named principal '{principal}'.",
file=sys.stderr,
)
raise SystemExit(75)
if not title.isprintable() or re.match(r"^[A-Za-z-]+-[Bb]y:", title):
print(
f"BLOCK: PR title is not one printable, non-trailer line; refusing merge; "
f"escalate to named principal '{principal}'.",
file=sys.stderr,
)
raise SystemExit(75)
print(json.dumps({
"MergeTitleField": title,
"MergeMessageField": "\n".join(trailers),
}, separators=(",", ":")))
PY
}
merge_gitea_api_attempt() {
local host="$1" auth_mode="$2" credential="$3"
local api_url attempt_dir body_file raw_code commits_file fields_file context_file head_file payload_file work_root attempt_rc auth_config curl_rc
LAST_GITEA_HTTP_CODE="000"
LAST_GITEA_ERROR=""
api_url="https://${host}/api/v1/repos/${OWNER}/${REPO}/pulls/${PR_NUMBER}/merge"
work_root="${AGENT_WORK_ROOT:-${HOME:-/tmp}/mosaic/agent-work}"
mkdir -p "$work_root"
attempt_dir=$(mktemp -d "$work_root/pr-merge-attempt.XXXXXX")
chmod 0700 "$attempt_dir"
MERGE_TEMP_DIRS+=("$attempt_dir")
body_file=$(mktemp "$attempt_dir/api-response.XXXXXX")
fields_file=$(mktemp "$attempt_dir/message-fields.XXXXXX")
payload_file=$(mktemp "$attempt_dir/payload.XXXXXX")
printf '{}' > "$fields_file"
if [[ "$CO_AUTHOR_TRAILERS" == true ]]; then
commits_file=$(mktemp "$attempt_dir/pr-merge-commits-input.XXXXXX")
context_file=$(mktemp "$attempt_dir/pr-merge-message-context.XXXXXX")
head_file=$(mktemp "$attempt_dir/pr-merge-head-input.XXXXXX")
printf '%s\0%s\0%s\0' "$PR_AUTHOR" "$PR_TITLE" "$ESCALATE_TO" > "$context_file"
if fetch_gitea_pr_head "$host" "$auth_mode" "$credential" "$attempt_dir" > "$head_file"; then
:
else
attempt_rc=$?
rm -f "$body_file" "$fields_file" "$payload_file" "$commits_file" "$context_file" "$head_file"
return "$attempt_rc"
fi
if [[ "$(<"$head_file")" != "$HEAD_SHA" ]]; then
echo "BLOCK: authenticated PR head moved from reviewed $HEAD_SHA to $(<"$head_file"); refusing merge; escalate to named principal '$ESCALATE_TO'." >&2
rm -f "$body_file" "$fields_file" "$payload_file" "$commits_file" "$context_file" "$head_file"
return 75
fi
if fetch_gitea_pr_commits "$host" "$auth_mode" "$credential" "$attempt_dir" > "$commits_file"; then
:
else
attempt_rc=$?
rm -f "$body_file" "$fields_file" "$payload_file" "$commits_file" "$context_file" "$head_file"
return "$attempt_rc"
fi
if build_coauthor_message_fields "$commits_file" "$context_file" "$head_file" > "$fields_file"; then
:
else
attempt_rc=$?
rm -f "$body_file" "$fields_file" "$payload_file" "$commits_file" "$context_file" "$head_file"
return "$attempt_rc"
fi
rm -f "$commits_file" "$context_file" "$head_file"
fi
if ! python3 - "$fields_file" "$HEAD_SHA" "$DELETE_BRANCH" > "$payload_file" <<'PY'
import json
import sys
with open(sys.argv[1], encoding="utf-8") as handle:
fields = json.load(handle)
head_sha, delete_branch = sys.argv[2:]
payload = {"Do": "squash", "head_commit_id": head_sha} payload = {"Do": "squash", "head_commit_id": head_sha}
if delete_branch == "true": if delete_branch == "true":
payload["delete_branch_after_merge"] = True payload["delete_branch_after_merge"] = True
payload.update(fields)
allowed = {"Do", "head_commit_id", "delete_branch_after_merge", "MergeTitleField", "MergeMessageField"}
if payload.get("Do") != "squash" or set(payload) - allowed:
raise SystemExit(1)
print(json.dumps(payload, separators=(",", ":"))) print(json.dumps(payload, separators=(",", ":")))
PY PY
) then
rm -f "$body_file" "$fields_file" "$payload_file"
return 1
fi
rm -f "$fields_file"
token=$(get_gitea_token "$host" || true) if ! auth_config=$(write_curl_auth_config "$auth_mode" "$credential"); then
if [[ -n "$token" ]]; then echo "Error: Could not construct Gitea authentication config; refusing request." >&2
raw_code=$(curl -sS -w '%{http_code}' -o "$body_file" \ rm -f "$body_file" "$payload_file"
-X POST \ return 1
-H "User-Agent: curl/8" \ fi
-H "Authorization: token $token" \ raw_code=$(curl -sS -K - "${GITEA_CURL_BOUNDS[@]}" -w '%{http_code}' -o "$body_file" \
-X POST -H "User-Agent: curl/8" \
-H 'Content-Type: application/json' \ -H 'Content-Type: application/json' \
-d "$payload" \ --data-binary "@$payload_file" "$api_url" <<<"$auth_config")
"$api_url" || true) curl_rc=$?
if [[ "$raw_code" =~ ^2 ]]; then LAST_GITEA_HTTP_CODE="${raw_code:-000}"
rm -f "$body_file" if [[ "$curl_rc" -ne 0 ]]; then
return 0 LAST_GITEA_ERROR="curl transport failed (rc=$curl_rc)"
rm -f "$body_file" "$payload_file"
rm -rf -- "$attempt_dir"
return 1
fi fi
if [[ ! "$raw_code" =~ ^2 ]]; then
LAST_GITEA_ERROR=$(format_gitea_error_response "$body_file")
fi fi
rm -f "$body_file" "$payload_file"
rm -rf -- "$attempt_dir"
[[ "$raw_code" =~ ^2 ]]
}
basic_auth=$(get_gitea_basic_auth "$host" || true) merge_gitea_with_api() {
if [[ -n "$basic_auth" ]]; then local host="$1" token attempt_rc
raw_code=$(curl -sS -w '%{http_code}' -o "$body_file" \
-X POST \
-u "$basic_auth" \
-H "User-Agent: curl/8" \
-H 'Content-Type: application/json' \
-d "$payload" \
"$api_url" || true)
if [[ "$raw_code" =~ ^2 ]]; then
rm -f "$body_file"
return 0
fi
fi
python3 - "${raw_code:-000}" "$body_file" <<'PY' >&2 if ! token=$(get_gitea_token "$host"); then
import json echo "Error: Could not resolve the required Gitea token; refusing merge without changing principals." >&2
import sys return 1
code, path = sys.argv[1], sys.argv[2] fi
try: if [[ -z "$token" ]]; then
with open(path, encoding="utf-8", errors="replace") as handle: echo "Error: Required Gitea token resolved empty; refusing merge without changing principals." >&2
raw = handle.read(500) return 1
data = json.loads(raw) if raw else {} fi
message = data.get("message") or data.get("error") or raw or "empty response" if merge_gitea_api_attempt "$host" token "$token"; then
except Exception: return 0
try: else
message = open(path, encoding="utf-8", errors="replace").read(500) or "empty response" attempt_rc=$?
except Exception: fi
message = "unreadable response" if [[ "$attempt_rc" -eq 75 ]]; then
print(f"Error: Gitea API merge failed with HTTP {code}: {message}") return 75
PY fi
rm -f "$body_file" if [[ "$LAST_GITEA_HTTP_CODE" != "401" ]]; then
echo "Error: Gitea API merge failed with the identity-bound token (HTTP ${LAST_GITEA_HTTP_CODE:-000}).${LAST_GITEA_ERROR:+ Provider response: $LAST_GITEA_ERROR}" >&2
return 1
fi
echo "Error: Gitea API rejected the identity-bound token with HTTP 401; refusing cross-principal credential fallback." >&2
return 1 return 1
} }
@@ -195,11 +602,10 @@ if [[ "$DRY_RUN" == true ]]; then
echo "Error: Cannot determine host from origin remote URL" >&2 echo "Error: Cannot determine host from origin remote URL" >&2
exit 1 exit 1
} }
TEA_LOGIN="$(get_gitea_login_for_host "$HOST" || true)" if [[ "$CO_AUTHOR_TRAILERS" == true ]]; then
if [[ -n "$TEA_LOGIN" ]]; then echo "Dry run: would verify PR commit authors and merge PR #$PR_NUMBER on $HOST with authenticated Gitea API message fields (base=$BASE_BRANCH, method=squash)."
echo "Dry run: would merge PR #$PR_NUMBER on $HOST with tea login '$TEA_LOGIN' (base=$BASE_BRANCH, method=squash)."
else else
echo "Dry run: would merge PR #$PR_NUMBER on $HOST with authenticated Gitea API fallback (base=$BASE_BRANCH, method=squash)." echo "Dry run: would merge PR #$PR_NUMBER on $HOST with the authenticated exact-head Gitea API path (base=$BASE_BRANCH, method=squash)."
fi fi
else else
echo "Dry run: would merge PR #$PR_NUMBER on $PLATFORM (base=$BASE_BRANCH, method=squash)." echo "Dry run: would merge PR #$PR_NUMBER on $PLATFORM (base=$BASE_BRANCH, method=squash)."
@@ -209,6 +615,10 @@ fi
case "$PLATFORM" in case "$PLATFORM" in
github) github)
if [[ "$CO_AUTHOR_TRAILERS" == true ]]; then
echo "Error: --co-author-trailers currently requires the Gitea REST message-field contract." >&2
exit 1
fi
cmd=(gh pr merge "$PR_NUMBER" --squash --match-head-commit "$HEAD_SHA") cmd=(gh pr merge "$PR_NUMBER" --squash --match-head-commit "$HEAD_SHA")
[[ "$DELETE_BRANCH" == true ]] && cmd+=(--delete-branch) [[ "$DELETE_BRANCH" == true ]] && cmd+=(--delete-branch)
"${cmd[@]}" "${cmd[@]}"
@@ -219,7 +629,7 @@ case "$PLATFORM" in
exit 1 exit 1
} }
# Gitea's API head_commit_id is an atomic compare-and-merge precondition. # Gitea's API head_commit_id is an atomic compare-and-merge precondition.
# tea cannot express it, so exact-head merges use the authenticated API path. # tea cannot express it, so every Gitea merge uses the authenticated API path.
merge_gitea_with_api "$HOST" merge_gitea_with_api "$HOST"
;; ;;
*) *)
@@ -43,22 +43,22 @@ SH
chmod +x "$STUB_DIR/gh" chmod +x "$STUB_DIR/gh"
run_guard() { run_guard() {
local mode="$1" local mode="$1" purpose="${2:-push}"
( (
cd "$REPO_DIR" || exit cd "$REPO_DIR" || exit
export PATH="$STUB_DIR:$PATH" export PATH="$STUB_DIR:$PATH"
export MOSAIC_GH_CHECK_MODE="$mode" export MOSAIC_GH_CHECK_MODE="$mode"
export MOSAIC_GH_CALL_LOG="$WORK_DIR/gh-calls.log" export MOSAIC_GH_CALL_LOG="$WORK_DIR/gh-calls.log"
export MOSAIC_CI_QUEUE_AUDIT_LOG="$WORK_DIR/audit.jsonl" export MOSAIC_CI_QUEUE_AUDIT_LOG="$WORK_DIR/audit.jsonl"
"$SCRIPT_DIR/ci-queue-wait.sh" --purpose push -t 0 -i 0 "$SCRIPT_DIR/ci-queue-wait.sh" --purpose "$purpose" -t 0 -i 0
) )
} }
failures=0 failures=0
assert_case() { assert_case() {
local mode="$1" expected_rc="$2" expected_state="$3" output rc local mode="$1" expected_rc="$2" expected_state="$3" purpose="${4:-push}" output rc
set +e set +e
output=$(run_guard "$mode" 2>&1) output=$(run_guard "$mode" "$purpose" 2>&1)
rc=$? rc=$?
set -e set -e
if [[ "$expected_rc" == zero && "$rc" -ne 0 ]]; then if [[ "$expected_rc" == zero && "$rc" -ne 0 ]]; then
@@ -79,10 +79,12 @@ set -e
: > "$WORK_DIR/gh-calls.log" : > "$WORK_DIR/gh-calls.log"
assert_case success zero terminal-success assert_case success zero terminal-success
assert_case pending nonzero pending assert_case pending nonzero pending
assert_case failure nonzero terminal-failure assert_case failure zero terminal-failure
assert_case late-failure nonzero terminal-failure assert_case late-failure zero terminal-failure
assert_case failure nonzero terminal-failure merge
assert_case late-failure nonzero terminal-failure merge
if [[ $(grep -c 'check-runs?per_page=100&filter=latest' "$WORK_DIR/gh-calls.log") -lt 4 ]]; then if [[ $(grep -c 'check-runs?per_page=100&filter=latest' "$WORK_DIR/gh-calls.log") -lt 6 ]]; then
echo "FAIL: expected every case to query all Checks API pages" >&2 echo "FAIL: expected every case to query all Checks API pages" >&2
failures=$((failures + 1)) failures=$((failures + 1))
fi fi
@@ -92,4 +94,4 @@ if [[ "$failures" -ne 0 ]]; then
exit 1 exit 1
fi fi
echo "GitHub check-runs regression passed (4/4 cases, including later-page failure)" echo "GitHub check-runs regression passed (6/6 purpose-aware cases, including later-page failure)"
@@ -9,10 +9,51 @@ WORK_DIR="${MOSAIC_TEST_WORK_DIR:-$PWD/.mosaic-test-work/ci-queue-wait-tristate}
REPO_DIR="$WORK_DIR/repo" REPO_DIR="$WORK_DIR/repo"
STUB_DIR="$WORK_DIR/stubs" STUB_DIR="$WORK_DIR/stubs"
AUDIT_LOG="$WORK_DIR/audit/ci-queue-wait.jsonl" AUDIT_LOG="$WORK_DIR/audit/ci-queue-wait.jsonl"
STATUS_OBSERVED="$WORK_DIR/status-observed"
CLOCK_LOG="$WORK_DIR/clock.log"
WATCHDOG_PYTHON="/usr/bin/python3"
WATCHDOG_SCRIPT="$WORK_DIR/real-clock-watchdog.py"
WATCHDOG_TIMEOUT_SEC=5
WATCHDOG_EXIT=90
FEATURE_BRANCH="fix/rm-03-fixture" FEATURE_BRANCH="fix/rm-03-fixture"
if [[ ! -x "$WATCHDOG_PYTHON" ]]; then
echo "FAIL setup: required real-clock watchdog runtime is unavailable at $WATCHDOG_PYTHON" >&2
exit 1
fi
rm -rf "$WORK_DIR" rm -rf "$WORK_DIR"
mkdir -p "$REPO_DIR" "$STUB_DIR" mkdir -p "$REPO_DIR" "$STUB_DIR"
cat > "$WATCHDOG_SCRIPT" <<'PY'
import os
import signal
import subprocess
import sys
if len(sys.argv) < 3:
raise SystemExit(2)
timeout_seconds = float(sys.argv[1])
process = subprocess.Popen(sys.argv[2:], start_new_session=True)
try:
return_code = process.wait(timeout=timeout_seconds)
except subprocess.TimeoutExpired:
try:
os.killpg(process.pid, signal.SIGKILL)
except ProcessLookupError:
pass
process.wait()
print(
f"FAIL HANG watchdog: subject exceeded {timeout_seconds:g}s "
"before completing its intended path",
file=sys.stderr,
)
raise SystemExit(90)
if return_code < 0:
raise SystemExit(128 - return_code)
raise SystemExit(return_code)
PY
git -C "$REPO_DIR" init -q git -C "$REPO_DIR" init -q
git -C "$REPO_DIR" checkout -q -b "$FEATURE_BRANCH" git -C "$REPO_DIR" checkout -q -b "$FEATURE_BRANCH"
git -C "$REPO_DIR" remote add origin https://git.example.test/acme/widgets.git git -C "$REPO_DIR" remote add origin https://git.example.test/acme/widgets.git
@@ -33,6 +74,9 @@ printf '%s\n' "$url" >> "${MOSAIC_STUB_URL_LOG:?}"
case "$url" in case "$url" in
*/branches/*) */branches/*)
if [[ "${MOSAIC_STUB_BRANCH_MODE:-ok}" == "hang-before-provider" ]]; then
while :; do :; done
fi
if [[ "${MOSAIC_STUB_BRANCH_MODE:-ok}" == "unreachable" ]]; then if [[ "${MOSAIC_STUB_BRANCH_MODE:-ok}" == "unreachable" ]]; then
exit 7 exit 7
fi fi
@@ -44,6 +88,7 @@ case "$url" in
fi fi
;; ;;
*/status) */status)
: > "${MOSAIC_STUB_STATUS_OBSERVED:?}"
case "${MOSAIC_STUB_STATUS_MODE:?}" in case "${MOSAIC_STUB_STATUS_MODE:?}" in
success) printf '%s' '{"state":"success","statuses":[{"status":"success"}]}' ;; success) printf '%s' '{"state":"success","statuses":[{"status":"success"}]}' ;;
pending) printf '%s' '{"state":"pending","statuses":[{"status":"pending","context":"ci/test"}]}' ;; pending) printf '%s' '{"state":"pending","statuses":[{"status":"pending","context":"ci/test"}]}' ;;
@@ -53,6 +98,7 @@ case "$url" in
malformed) printf '%s' 'not-json' ;; malformed) printf '%s' 'not-json' ;;
malformed-statuses-type) printf '%s' '{"state":"success","statuses":"corrupt"}' ;; malformed-statuses-type) printf '%s' '{"state":"success","statuses":"corrupt"}' ;;
malformed-status-entry) printf '%s' '{"state":"success","statuses":[null]}' ;; malformed-status-entry) printf '%s' '{"state":"success","statuses":[null]}' ;;
unknown) printf '%s' '{"state":"success","statuses":[{"status":"cancelled"}]}' ;;
large-success) large-success)
python3 -c 'import json; print(json.dumps({"state":"success", "statuses":[{"status":"success"}], "padding":"x" * (160 * 1024)}), end="")' python3 -c 'import json; print(json.dumps({"state":"success", "statuses":[{"status":"success"}], "padding":"x" * (160 * 1024)}), end="")'
;; ;;
@@ -63,7 +109,31 @@ case "$url" in
*) echo "unexpected curl URL: $url" >&2; exit 2 ;; *) echo "unexpected curl URL: $url" >&2; exit 2 ;;
esac esac
SH SH
chmod +x "$STUB_DIR/curl"
cat > "$STUB_DIR/date" <<'SH'
#!/usr/bin/env bash
set -euo pipefail
if [[ "$#" -ne 1 || "$1" != "+%s" ]]; then
echo "unexpected date invocation: $*" >&2
exit 2
fi
if [[ -e "${MOSAIC_STUB_STATUS_OBSERVED:?}" ]]; then
printf 'date-phase=after-status\n' >> "${MOSAIC_STUB_CLOCK_LOG:?}"
printf '1002\n'
else
printf 'date-phase=before-status\n' >> "${MOSAIC_STUB_CLOCK_LOG:?}"
printf '1000\n'
fi
SH
cat > "$STUB_DIR/sleep" <<'SH'
#!/usr/bin/env bash
set -euo pipefail
printf 'sleep-after-status=%s\n' "$*" >> "${MOSAIC_STUB_CLOCK_LOG:?}"
SH
chmod +x "$STUB_DIR/curl" "$STUB_DIR/date" "$STUB_DIR/sleep"
run_guard() { run_guard() {
local status_mode="$1" local status_mode="$1"
@@ -83,13 +153,46 @@ run_guard() {
export GITEA_URL=https://git.example.test export GITEA_URL=https://git.example.test
export MOSAIC_STUB_STATUS_MODE="$status_mode" export MOSAIC_STUB_STATUS_MODE="$status_mode"
fi fi
rm -f "$STATUS_OBSERVED" "$CLOCK_LOG"
export MOSAIC_STUB_URL_LOG="$WORK_DIR/urls.log" export MOSAIC_STUB_URL_LOG="$WORK_DIR/urls.log"
export MOSAIC_STUB_STATUS_OBSERVED="$STATUS_OBSERVED"
export MOSAIC_STUB_CLOCK_LOG="$CLOCK_LOG"
export MOSAIC_CI_QUEUE_AUDIT_LOG="$audit_log" export MOSAIC_CI_QUEUE_AUDIT_LOG="$audit_log"
"$SCRIPT_DIR/ci-queue-wait.sh" --purpose "${MOSAIC_TEST_PURPOSE:-push}" -t 0 -i 0 "$@" # Provider observation is the synchronization event. The one-second
# timeout is subject semantics under virtual time, never a wall wait.
# The absolute Python runtime uses an internal monotonic wait and kills
# the subject's isolated process group. Neither operation can resolve
# to the virtual date/sleep stubs at the front of PATH.
local subject_rc
if "$WATCHDOG_PYTHON" "$WATCHDOG_SCRIPT" "$WATCHDOG_TIMEOUT_SEC" \
"$SCRIPT_DIR/ci-queue-wait.sh" --purpose "${MOSAIC_TEST_PURPOSE:-push}" -t 1 -i 1 "$@"; then
subject_rc=0
else
subject_rc=$?
fi
return "$subject_rc"
) )
} }
failures=0 failures=0
assert_provider_observed() {
local name="$1" require_expiration="${2:-0}"
if [[ ! -e "$STATUS_OBSERVED" ]]; then
echo "FAIL $name: status provider was not observed" >&2
failures=$((failures + 1))
fi
if [[ ! -s "$CLOCK_LOG" ]] || ! grep -q '^date-phase=before-status$' "$CLOCK_LOG"; then
echo "FAIL $name: virtual clock interception did not run before provider observation" >&2
failures=$((failures + 1))
fi
if [[ "$require_expiration" -eq 1 ]]; then
if ! grep -q '^sleep-after-status=' "$CLOCK_LOG" || ! grep -q '^date-phase=after-status$' "$CLOCK_LOG"; then
echo "FAIL $name: pending path did not expire after provider observation" >&2
failures=$((failures + 1))
fi
fi
}
run_assertion() { run_assertion() {
local name="$1" expected_rc="$2" status_mode="$3" required_text="$4" local name="$1" expected_rc="$2" status_mode="$3" required_text="$4"
local output rc local output rc
@@ -124,22 +227,59 @@ run_assertion() {
printf '%s\n' "$output" >&2 printf '%s\n' "$output" >&2
failures=$((failures + 1)) failures=$((failures + 1))
fi fi
if [[ "$status_mode" != "credential-unresolvable" ]]; then
if [[ "$status_mode" == "pending" ]]; then
assert_provider_observed "$name" 1
else
assert_provider_observed "$name"
fi
fi
} }
set -e set -e
: > "$WORK_DIR/urls.log" : > "$WORK_DIR/urls.log"
run_assertion success zero success 'state=terminal-success' # Push readiness is queue clearance, not proof that prior CI succeeded.
run_assertion pending nonzero pending 'ASSERTED_NOT_READY' run_assertion push-success zero success 'state=terminal-success'
run_assertion failure nonzero failure 'ASSERTED_NOT_READY' run_assertion push-pending nonzero pending 'ASSERTED_NOT_READY'
run_assertion no-status nonzero no-status 'ASSERTED_NOT_READY' run_assertion push-failure zero failure 'queue-clear state=terminal-failure purpose=push'
run_assertion aggregate-success-no-status nonzero aggregate-success-no-status 'ASSERTED_NOT_READY' run_assertion push-no-status zero no-status 'queue-clear state=no-status purpose=push'
run_assertion malformed nonzero malformed 'ASSERTED_NOT_READY' run_assertion push-aggregate-success-no-status zero aggregate-success-no-status 'queue-clear state=no-status purpose=push'
run_assertion malformed-statuses-type nonzero malformed-statuses-type 'ASSERTED_NOT_READY' run_assertion push-require-status nonzero no-status 'ASSERTED_NOT_READY state=no-status' --require-status
run_assertion malformed-status-entry nonzero malformed-status-entry 'ASSERTED_NOT_READY' run_assertion push-malformed nonzero malformed 'ASSERTED_NOT_READY'
run_assertion large-payload not126 large-success 'state=terminal-success' run_assertion push-malformed-statuses-type nonzero malformed-statuses-type 'ASSERTED_NOT_READY'
run_assertion push-malformed-status-entry nonzero malformed-status-entry 'ASSERTED_NOT_READY'
run_assertion push-unknown nonzero unknown 'ASSERTED_NOT_READY'
run_assertion push-large-payload not126 large-success 'state=terminal-success'
run_assertion credential-unresolvable zero credential-unresolvable 'CANNOT_ASSERT' run_assertion credential-unresolvable zero credential-unresolvable 'CANNOT_ASSERT'
run_assertion provider-unreachable zero unreachable 'CANNOT_ASSERT' run_assertion provider-unreachable zero unreachable 'CANNOT_ASSERT'
# Merge readiness remains fail-closed and requires exact-head terminal success.
MOSAIC_TEST_PURPOSE=merge run_assertion merge-success zero success 'state=terminal-success'
MOSAIC_TEST_PURPOSE=merge run_assertion merge-failure nonzero failure 'ASSERTED_NOT_READY state=terminal-failure'
MOSAIC_TEST_PURPOSE=merge run_assertion merge-no-status nonzero no-status 'ASSERTED_NOT_READY state=no-status'
MOSAIC_TEST_PURPOSE=merge run_assertion merge-unknown nonzero unknown 'ASSERTED_NOT_READY state=unknown'
# Positive liveness control: a subject mutant hangs before the branch lookup
# can reach the status provider. Only the independent real-clock watchdog may
# terminate it, and its failure must be distinct from subject timeout rc=124.
set +e
watchdog_output=$(MOSAIC_STUB_BRANCH_MODE=hang-before-provider run_guard success "$AUDIT_LOG" 2>&1)
watchdog_rc=$?
set -e
if [[ "$watchdog_rc" -ne "$WATCHDOG_EXIT" ]]; then
echo "FAIL watchdog-control: expected hang-specific rc=$WATCHDOG_EXIT, got rc=$watchdog_rc" >&2
failures=$((failures + 1))
fi
if [[ "$watchdog_output" != *"FAIL HANG watchdog:"* ]]; then
echo "FAIL watchdog-control: expected distinct hang-specific diagnostic" >&2
printf '%s\n' "$watchdog_output" >&2
failures=$((failures + 1))
fi
if [[ -e "$STATUS_OBSERVED" ]]; then
echo "FAIL watchdog-control: hanging mutant unexpectedly reached the status provider" >&2
failures=$((failures + 1))
fi
if [[ ! -s "$AUDIT_LOG" ]] || ! grep -q '"outcome":"CANNOT_ASSERT"' "$AUDIT_LOG"; then if [[ ! -s "$AUDIT_LOG" ]] || ! grep -q '"outcome":"CANNOT_ASSERT"' "$AUDIT_LOG"; then
echo "FAIL provider-unreachable-audit: expected durable CANNOT_ASSERT JSONL record" >&2 echo "FAIL provider-unreachable-audit: expected durable CANNOT_ASSERT JSONL record" >&2
failures=$((failures + 1)) failures=$((failures + 1))
@@ -160,6 +300,7 @@ if [[ "$merge_unreachable_output" != *"CANNOT_ASSERT"* ]]; then
echo "FAIL merge-provider-unreachable: expected loud CANNOT_ASSERT diagnostic" >&2 echo "FAIL merge-provider-unreachable: expected loud CANNOT_ASSERT diagnostic" >&2
failures=$((failures + 1)) failures=$((failures + 1))
fi fi
assert_provider_observed merge-provider-unreachable
merge_audit_lines_after=$(wc -l < "$AUDIT_LOG") merge_audit_lines_after=$(wc -l < "$AUDIT_LOG")
if [[ "$merge_audit_lines_after" -le "$merge_audit_lines_before" ]]; then if [[ "$merge_audit_lines_after" -le "$merge_audit_lines_before" ]]; then
echo "FAIL merge-provider-unreachable: expected an additional audit record" >&2 echo "FAIL merge-provider-unreachable: expected an additional audit record" >&2
@@ -223,6 +364,7 @@ if [[ "$audit_failure_output" != *"audit"* ]]; then
echo "FAIL audit-unavailable: expected loud audit failure diagnostic" >&2 echo "FAIL audit-unavailable: expected loud audit failure diagnostic" >&2
failures=$((failures + 1)) failures=$((failures + 1))
fi fi
assert_provider_observed audit-unavailable
if [[ "$failures" -ne 0 ]]; then if [[ "$failures" -ne 0 ]]; then
echo "ci-queue-wait tri-state regression failed ($failures assertions)" >&2 echo "ci-queue-wait tri-state regression failed ($failures assertions)" >&2
@@ -51,22 +51,23 @@ for arg in "$@"; do
prev="" prev=""
continue continue
fi fi
if [[ "$prev" == "-d" ]]; then if [[ "$prev" == "data" ]]; then
post_data="$arg" post_data="$arg"
[[ "$post_data" == @* ]] && post_data=$(<"${post_data#@}")
prev="" prev=""
continue continue
fi fi
if [[ "$arg" == "-o" ]]; then if [[ "$prev" == "config" ]]; then
prev="-o" [[ "$arg" == "-" ]] && cat >/dev/null
prev=""
continue continue
fi fi
if [[ "$arg" == "-d" ]]; then case "$arg" in
prev="-d" -o) prev="-o" ;;
continue -d|--data|--data-binary) prev="data" ;;
fi -K|--config) prev="config" ;;
if [[ "$arg" == "-w" ]]; then -w) write_code=true ;;
write_code=true esac
fi
done done
emit_response() { emit_response() {
local body="$1" local body="$1"
@@ -17,9 +17,10 @@ make_fixture() {
cp "$SCRIPT_DIR/detect-platform.sh" "$tools/detect-platform.sh" cp "$SCRIPT_DIR/detect-platform.sh" "$tools/detect-platform.sh"
git -C "$root/repo" init -q git -C "$root/repo" init -q
git -C "$root/repo" remote add origin "$remote" git -C "$root/repo" remote add origin "$remote"
local base_branch="${3:-main}"
cat > "$tools/pr-metadata.sh" <<SH cat > "$tools/pr-metadata.sh" <<SH
#!/usr/bin/env bash #!/usr/bin/env bash
printf '%s\n' '{"baseRefName":"main","headRefName":"fix/pinned","headRefOid":"$SHA","headRepository":"contributor/widgets-fork"}' printf '%s\n' '{"baseRefName":"$base_branch","headRefName":"fix/pinned","headRefOid":"$SHA","headRepository":"contributor/widgets-fork"}'
SH SH
cat > "$tools/ci-queue-wait.sh" <<'SH' cat > "$tools/ci-queue-wait.sh" <<'SH'
#!/usr/bin/env bash #!/usr/bin/env bash
@@ -29,20 +30,37 @@ SH
} }
rm -rf "$WORK_DIR" rm -rf "$WORK_DIR"
make_fixture gitea https://git.example.test/acme/widgets.git make_fixture gitea https://git.example.test/acme/widgets.git next
make_fixture github https://github.com/acme/widgets.git make_fixture github https://github.com/acme/widgets.git main
cat > "$WORK_DIR/gitea/curl" <<'SH' cat > "$WORK_DIR/gitea/curl" <<'SH'
#!/usr/bin/env bash #!/usr/bin/env bash
set -euo pipefail set -euo pipefail
payload="" payload=""
for ((i=1; i<=$#; i++)); do out_file=""
if [[ "${!i}" == "-d" ]]; then while [[ $# -gt 0 ]]; do
j=$((i + 1)) case "$1" in
payload="${!j}" -d|--data|--data-binary)
fi payload="$2"
[[ "$payload" == @* ]] && payload=$(<"${payload#@}")
shift 2
;;
-o)
out_file="$2"
shift 2
;;
-K|--config)
[[ "$2" == "-" ]] && cat >/dev/null
shift 2
;;
-w|-X|-H)
shift 2
;;
*) shift ;;
esac
done done
printf '%s' "$payload" > "${MOSAIC_MERGE_PAYLOAD_LOG:?}" printf '%s' "$payload" > "${MOSAIC_MERGE_PAYLOAD_LOG:?}"
[[ -n "$out_file" ]] && printf '{}' > "$out_file"
printf '200' printf '200'
SH SH
chmod +x "$WORK_DIR/gitea/curl" chmod +x "$WORK_DIR/gitea/curl"
@@ -0,0 +1,541 @@
#!/usr/bin/env bash
# Regression harness for the optional, identity-checked Gitea squash message.
set -u
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
SUBJECT="${MOSAIC_TEST_SUBJECT:-$SCRIPT_DIR/pr-merge.sh}"
WORK_DIR="${MOSAIC_TEST_WORK_DIR:-$PWD/.mosaic-test-work/pr-merge-message-field}"
ORIG_PATH="$PATH"
failures=0
rm -rf "$WORK_DIR"
mkdir -p "$WORK_DIR"
fail() {
echo "FAIL $1" >&2
failures=$((failures + 1))
}
make_case() {
local name="$1" case_dir
case_dir="$WORK_DIR/$name"
mkdir -p "$case_dir/bin" "$case_dir/agent"
cp "$SUBJECT" "$case_dir/pr-merge.sh"
chmod +x "$case_dir/pr-merge.sh"
cat > "$case_dir/detect-platform.sh" <<'SH'
#!/usr/bin/env bash
detect_platform() { PLATFORM=gitea; printf 'gitea\n'; }
get_repo_owner() { printf 'acme\n'; }
get_repo_name() { printf 'widgets\n'; }
get_remote_host() { printf 'git.example.test\n'; }
get_gitea_token() {
printf 'resolved\n' >> "${MOSAIC_TEST_TOKEN_RESOLUTION_LOG:?}"
if [[ "${MOSAIC_TEST_TOKEN_AVAILABLE:-true}" != "true" ]]; then
return 1
fi
printf 'fixture-token\n'
}
get_gitea_basic_auth() {
printf 'resolved\n' >> "${MOSAIC_TEST_BASIC_RESOLUTION_LOG:?}"
if [[ "${MOSAIC_TEST_BASIC_AVAILABLE:-false}" == "true" ]]; then
printf 'fixture-user:fixture-password\n'
return "${MOSAIC_TEST_BASIC_RC:-0}"
fi
return 1
}
get_gitea_login_for_host() { return 1; }
SH
cat > "$case_dir/pr-metadata.sh" <<'SH'
#!/usr/bin/env bash
if [[ "${MOSAIC_TEST_TITLE_MODE:-safe}" == "injection" ]]; then
title='Preserve authors\n\nCo-authored-by: victim <[email protected]>'
else
title='Preserve both branch authors'
fi
case "${MOSAIC_TEST_COMMITS_MODE:?}" in
verified) head_sha=2222222222222222222222222222222222222222 ;;
null-login|unsafe-identity) head_sha=3333333333333333333333333333333333333333 ;;
single) head_sha=1111111111111111111111111111111111111111 ;;
*) echo "unknown commits mode" >&2; exit 2 ;;
esac
printf '{"number":42,"title":"%s","author":"poster","baseRefName":"main","headRefName":"feature/fixture","headRefOid":"%s","headRepository":"acme/widgets"}\n' "$title" "$head_sha"
SH
cat > "$case_dir/ci-queue-wait.sh" <<'SH'
#!/usr/bin/env bash
exit 0
SH
cat > "$case_dir/bin/python3" <<'SH'
#!/usr/bin/env bash
for arg in "$@"; do
case "$arg" in
*"Preserve both branch authors"*|*"[email protected]"*)
: > "${MOSAIC_TEST_METADATA_ARGV_MARKER:?}"
;;
esac
done
exec "${MOSAIC_TEST_REAL_PYTHON:?}" "$@"
SH
cat > "$case_dir/bin/curl" <<'SH'
#!/usr/bin/env bash
set -eu
for arg in "$@"; do
case "$arg" in
*"Preserve both branch authors"*|*"[email protected]"*)
: > "${MOSAIC_TEST_METADATA_ARGV_MARKER:?}"
;;
esac
done
url=""
method="GET"
out_file=""
data=""
config=""
auth_mode="none"
has_max_filesize=0
has_max_time=0
has_connect_timeout=0
while [[ $# -gt 0 ]]; do
case "$1" in
-o)
out_file="$2"
shift 2
;;
-w)
shift 2
;;
-X)
method="$2"
shift 2
;;
-d|--data|--data-binary)
data="$2"
if [[ "$data" == @* ]]; then
data=$(<"${data#@}")
fi
shift 2
;;
-K|--config)
if [[ "$2" == "-" ]]; then
config=$(cat)
fi
shift 2
;;
--max-filesize)
has_max_filesize=1
shift 2
;;
--max-time)
has_max_time=1
shift 2
;;
--connect-timeout)
has_connect_timeout=1
shift 2
;;
-H|--header|-u|--user)
if [[ "$2" == *"fixture-token"* ]]; then
: > "${MOSAIC_TEST_TOKEN_ARGV_MARKER:?}"
fi
if [[ "$2" == *"fixture-password"* ]]; then
: > "${MOSAIC_TEST_BASIC_ARGV_MARKER:?}"
fi
shift 2
;;
http://*|https://*)
url="$1"
shift
;;
*)
shift
;;
esac
done
if [[ "$config" == *"Authorization: token fixture-token"* ]]; then
auth_mode="token"
: > "${MOSAIC_TEST_AUTH_CONFIG_MARKER:?}"
elif [[ "$config" == *"user = \"fixture-user:fixture-password\""* ]]; then
auth_mode="basic"
: > "${MOSAIC_TEST_BASIC_CONFIG_MARKER:?}"
fi
printf '%s %s %s\n' "$method" "$auth_mode" "$url" >> "${MOSAIC_TEST_CURL_LOG:?}"
printf '%s:%s:%s\n' "$has_max_filesize" "$has_max_time" "$has_connect_timeout" >> "${MOSAIC_TEST_CURL_BOUNDS_LOG:?}"
case "$url" in
*/pulls/42)
case "${MOSAIC_TEST_COMMITS_MODE:?}" in
verified) head_sha=2222222222222222222222222222222222222222 ;;
null-login|unsafe-identity) head_sha=3333333333333333333333333333333333333333 ;;
single) head_sha=1111111111111111111111111111111111111111 ;;
*) echo "unknown commits mode" >&2; exit 2 ;;
esac
if [[ "${MOSAIC_TEST_HEAD_MODE:-stable}" == "moved" ]]; then
head_sha=4444444444444444444444444444444444444444
fi
body="{\"head\":{\"sha\":\"$head_sha\"}}"
code=200
if [[ "${MOSAIC_TEST_FALLBACK_MODE:-none}" == "inspection" && "$auth_mode" == "token" ]]; then
body='{"message":"token rejected"}'
code=401
fi
;;
*/pulls/42/commits*)
case "${MOSAIC_TEST_COMMITS_MODE:?}" in
verified)
if [[ "${MOSAIC_TEST_EMAIL_MODE:-safe}" == "escape" ]]; then
body='[{"sha":"2222222222222222222222222222222222222222","commit":{"author":{"name":"Alice","email":"alice+\u001b[[email protected]"}},"author":{"login":"alice"}},{"sha":"1111111111111111111111111111111111111111","commit":{"author":{"name":"Poster","email":"[email protected]"}},"author":{"login":"poster"}}]'
else
body='[{"sha":"2222222222222222222222222222222222222222","commit":{"author":{"name":"Alice","email":"[email protected]"}},"author":{"login":"alice"}},{"sha":"1111111111111111111111111111111111111111","commit":{"author":{"name":"Poster","email":"[email protected]"}},"author":{"login":"poster"}}]'
fi
;;
null-login)
body='[{"sha":"1111111111111111111111111111111111111111","commit":{"author":{"name":"Poster","email":"[email protected]"}},"author":{"login":"poster"}},{"sha":"3333333333333333333333333333333333333333","commit":{"author":{"name":"Unresolved Author","email":"[email protected]\n\u001b[31m"}},"author":null}]'
;;
unsafe-identity)
body='[{"sha":"unsafe\n\u001b[31m","commit":{"author":{"name":"Unsafe","email":"not-an-email"}},"author":{"login":"unsafe"}},{"sha":"3333333333333333333333333333333333333333","commit":{"author":{"name":"Poster","email":"[email protected]"}},"author":{"login":"poster"}}]'
;;
single)
body='[{"sha":"1111111111111111111111111111111111111111","commit":{"author":{"name":"Poster","email":"[email protected]"}},"author":{"login":"poster"}}]'
;;
*)
echo "unknown commits mode" >&2
exit 2
;;
esac
code=200
if [[ "${MOSAIC_TEST_FALLBACK_MODE:-none}" == "inspection" && "$auth_mode" == "token" ]]; then
body='{"message":"token rejected"}'
code=401
fi
;;
*/pulls/42/merge)
body='{}'
code=200
if [[ "${MOSAIC_TEST_FALLBACK_MODE:-none}" == "merge" && "$auth_mode" == "token" ]]; then
body='{"message":"token rejected"}'
code=401
elif [[ "${MOSAIC_TEST_FALLBACK_MODE:-none}" == "provider-error" ]]; then
body='{"message":"branch policy rejected\n\u001b[31m"}'
code=409
elif [[ "${MOSAIC_TEST_FALLBACK_MODE:-none}" == "forbidden" ]]; then
body='{"message":"permission denied"}'
code=403
else
printf '%s' "$data" > "${MOSAIC_TEST_MERGE_PAYLOAD:?}"
fi
;;
*/users/*)
body='{"message":"not found"}'
code=404
;;
*)
body='{"message":"unexpected URL"}'
code=500
;;
esac
if [[ -n "$out_file" ]]; then
printf '%s' "$body" > "$out_file"
else
printf '%s' "$body"
fi
printf '%s' "$code"
case "${MOSAIC_TEST_CURL_FAILURE:-none}" in
oversize) exit 63 ;;
stalled) exit 28 ;;
esac
SH
chmod +x "$case_dir/detect-platform.sh" "$case_dir/pr-metadata.sh" \
"$case_dir/ci-queue-wait.sh" "$case_dir/bin/curl" "$case_dir/bin/python3"
printf '%s\n' "$case_dir"
}
run_case() {
local case_dir="$1" mode="$2"
shift 2
MOSAIC_TEST_COMMITS_MODE="$mode" \
MOSAIC_TEST_CURL_LOG="$case_dir/curl.log" \
MOSAIC_TEST_CURL_BOUNDS_LOG="$case_dir/curl-bounds.log" \
MOSAIC_TEST_MERGE_PAYLOAD="$case_dir/merge-payload.json" \
MOSAIC_TEST_TOKEN_ARGV_MARKER="$case_dir/token-in-argv" \
MOSAIC_TEST_BASIC_ARGV_MARKER="$case_dir/basic-in-argv" \
MOSAIC_TEST_AUTH_CONFIG_MARKER="$case_dir/auth-via-config" \
MOSAIC_TEST_BASIC_CONFIG_MARKER="$case_dir/basic-via-config" \
MOSAIC_TEST_TOKEN_RESOLUTION_LOG="$case_dir/token-resolution.log" \
MOSAIC_TEST_BASIC_RESOLUTION_LOG="$case_dir/basic-resolution.log" \
MOSAIC_TEST_METADATA_ARGV_MARKER="$case_dir/metadata-in-argv" \
MOSAIC_TEST_REAL_PYTHON="$(command -v python3)" \
AGENT_WORK_ROOT="$case_dir/agent" \
PATH="$case_dir/bin:$ORIG_PATH" \
"$case_dir/pr-merge.sh" -n 42 "$@"
}
# Verified multi-author path: the non-poster trailer is built from one commit's
# linked author.login and that same commit's author email. No /users lookup.
verified_dir=$(make_case verified)
set +e
verified_output=$(run_case "$verified_dir" verified --co-author-trailers --escalate-to tl-mosaic 2>&1)
verified_rc=$?
set -e
if [[ "$verified_rc" -ne 0 ]]; then
fail "verified multi-author merge expected rc=0, got rc=$verified_rc: $verified_output"
elif [[ ! -s "$verified_dir/merge-payload.json" ]]; then
fail "verified multi-author merge did not reach the API payload"
else
python3 - "$verified_dir/merge-payload.json" <<'PY' || fail "verified payload did not preserve squash and exact message fields"
import json
import sys
payload = json.load(open(sys.argv[1], encoding="utf-8"))
assert payload == {
"Do": "squash",
"head_commit_id": "2222222222222222222222222222222222222222",
"MergeTitleField": "Preserve both branch authors",
"MergeMessageField": "Co-authored-by: alice <[email protected]>",
}, payload
PY
fi
[[ -e "$verified_dir/auth-via-config" ]] || fail "verified path did not authenticate curl through stdin config"
[[ ! -e "$verified_dir/token-in-argv" ]] || fail "verified path placed the Gitea token in curl argv"
[[ ! -e "$verified_dir/metadata-in-argv" ]] || fail "verified path placed PR title or contributor email in child argv"
[[ "$(wc -l < "$verified_dir/token-resolution.log")" -eq 1 ]] || fail "verified path did not bind inspection and merge to one credential resolution"
if grep -q '/users/' "$verified_dir/curl.log" 2>/dev/null; then
fail "verified path performed a forbidden second /users lookup"
fi
if grep -qv '^1:1:1$' "$verified_dir/curl-bounds.log"; then
fail "verified path did not apply size/max-time/connect-time bounds to every provider download"
fi
# A linked email containing a terminal escape must block before mutation.
escape_email_dir=$(make_case escape-email)
set +e
escape_email_output=$(MOSAIC_TEST_EMAIL_MODE=escape run_case "$escape_email_dir" verified --co-author-trailers --escalate-to tl-mosaic 2>&1)
escape_email_rc=$?
set -e
[[ "$escape_email_rc" -ne 0 ]] || fail "control-byte email unexpectedly passed"
[[ "$escape_email_output" == *"unusable linked identity"* ]] || fail "control-byte email refusal lost its diagnostic"
[[ ! -e "$escape_email_dir/merge-payload.json" ]] || fail "control-byte email reached the merge API"
# Curl transfer and duration failures must remain failures even with HTTP 200.
for failure_mode in oversize stalled; do
failure_dir=$(make_case "curl-$failure_mode")
set +e
failure_output=$(MOSAIC_TEST_CURL_FAILURE="$failure_mode" run_case "$failure_dir" verified --co-author-trailers --escalate-to tl-mosaic 2>&1)
failure_rc=$?
set -e
[[ "$failure_rc" -ne 0 ]] || fail "curl $failure_mode failure was discarded: $failure_output"
[[ ! -e "$failure_dir/merge-payload.json" ]] || fail "curl $failure_mode failure reached the merge API"
done
# The authenticated head is re-read under the mutation credential but cannot
# replace the canonical preflight/review head. A move blocks before enumeration
# or mutation even though the provider returned a valid new SHA.
moved_dir=$(make_case moved-head)
set +e
moved_output=$(MOSAIC_TEST_HEAD_MODE=moved \
run_case "$moved_dir" verified --co-author-trailers --escalate-to tl-mosaic 2>&1)
moved_rc=$?
set -e
[[ "$moved_rc" -ne 0 ]] || fail "moved authenticated head unexpectedly passed"
[[ "$moved_output" == *"authenticated PR head moved from reviewed"* ]] || fail "moved head refusal lost its diagnostic"
[[ "$moved_output" == *"tl-mosaic"* ]] || fail "moved head refusal omitted the named escalation principal"
[[ ! -e "$moved_dir/merge-payload.json" ]] || fail "moved head refusal reached the merge API"
moved_sequence=$(awk '{print $1 ":" $2}' "$moved_dir/curl.log" | paste -sd, -)
[[ "$moved_sequence" == "GET:token" ]] || fail "moved head refusal performed post-move inspection/mutation (calls=$moved_sequence)"
# Token resolution failure is not an authentication response. It must fail
# closed instead of borrowing a Basic credential under a different principal.
token_missing_dir=$(make_case token-missing)
set +e
token_missing_output=$(MOSAIC_TEST_TOKEN_AVAILABLE=false MOSAIC_TEST_BASIC_AVAILABLE=true \
run_case "$token_missing_dir" single 2>&1)
token_missing_rc=$?
set -e
[[ "$token_missing_rc" -ne 0 ]] || fail "missing token unexpectedly borrowed Basic Auth"
[[ "$token_missing_output" == *"required Gitea token"* ]] || fail "missing token refusal lost its diagnostic"
[[ ! -e "$token_missing_dir/basic-resolution.log" ]] || fail "missing token resolved Basic Auth after identity failure"
[[ ! -e "$token_missing_dir/curl.log" ]] || fail "missing token reached a provider request"
# A failed Basic resolver must never use its nonempty output or reach mutation.
basic_rc_dir=$(make_case basic-resolver-rc)
set +e
basic_rc_output=$(MOSAIC_TEST_BASIC_AVAILABLE=true MOSAIC_TEST_BASIC_RC=91 MOSAIC_TEST_FALLBACK_MODE=inspection \
run_case "$basic_rc_dir" verified --co-author-trailers --escalate-to tl-mosaic 2>&1)
basic_rc_rc=$?
set -e
[[ "$basic_rc_rc" -ne 0 ]] || fail "failed Basic resolver output unexpectedly authorized a merge: $basic_rc_output"
[[ ! -e "$basic_rc_dir/merge-payload.json" ]] || fail "failed Basic resolver reached the merge API"
# HTTP 401 never changes principals: inspection rejection fails closed without
# resolving or attempting Basic Auth.
fallback_inspect_dir=$(make_case fallback-inspection)
set +e
fallback_inspect_output=$(MOSAIC_TEST_BASIC_AVAILABLE=true MOSAIC_TEST_FALLBACK_MODE=inspection \
run_case "$fallback_inspect_dir" verified --co-author-trailers --escalate-to tl-mosaic 2>&1)
fallback_inspect_rc=$?
set -e
[[ "$fallback_inspect_rc" -ne 0 ]] || fail "inspection token rejection unexpectedly changed principals"
[[ "$fallback_inspect_output" == *"refusing cross-principal credential fallback"* ]] || fail "inspection token rejection lost its refusal diagnostic"
[[ ! -e "$fallback_inspect_dir/basic-resolution.log" ]] || fail "inspection token rejection resolved Basic Auth"
[[ ! -e "$fallback_inspect_dir/merge-payload.json" ]] || fail "inspection token rejection reached merge mutation"
inspect_sequence=$(awk '{print $1 ":" $2}' "$fallback_inspect_dir/curl.log" | paste -sd, -)
[[ "$inspect_sequence" == "GET:token" ]] || fail "inspection rejection made unexpected provider calls (calls=$inspect_sequence)"
# Token rejection at merge likewise fails closed without cross-principal retry.
fallback_merge_dir=$(make_case fallback-merge)
set +e
fallback_merge_output=$(MOSAIC_TEST_BASIC_AVAILABLE=true MOSAIC_TEST_FALLBACK_MODE=merge \
run_case "$fallback_merge_dir" verified --co-author-trailers --escalate-to tl-mosaic 2>&1)
fallback_merge_rc=$?
set -e
[[ "$fallback_merge_rc" -ne 0 ]] || fail "merge token rejection unexpectedly changed principals"
[[ "$fallback_merge_output" == *"refusing cross-principal credential fallback"* ]] || fail "merge token rejection lost its refusal diagnostic"
[[ ! -e "$fallback_merge_dir/basic-resolution.log" ]] || fail "merge token rejection resolved Basic Auth"
[[ ! -e "$fallback_merge_dir/merge-payload.json" ]] || fail "merge token rejection recorded a successful payload"
merge_sequence=$(awk '{print $1 ":" $2}' "$fallback_merge_dir/curl.log" | paste -sd, -)
[[ "$merge_sequence" == "GET:token,GET:token,POST:token" ]] || fail "merge rejection made unexpected provider calls (calls=$merge_sequence)"
# BLOCK path: a commit email exists but author.login is null. It must name both
# facts, name the escalation principal, and never reach the merge endpoint.
null_dir=$(make_case null-login)
set +e
null_output=$(run_case "$null_dir" null-login --co-author-trailers --escalate-to tl-mosaic 2>&1)
null_rc=$?
set -e
[[ "$null_rc" -ne 0 ]] || fail "null-login author expected a non-zero BLOCK"
[[ "$null_output" == *"BLOCK"* ]] || fail "null-login author omitted BLOCK diagnostic"
[[ "$null_output" == *"author.login=NULL"* ]] || fail "null-login author omitted the null provider fact"
[[ "$null_output" == *"[email protected]"* ]] || fail "null-login author omitted the commit email fact"
[[ "$null_output" == *'\n\x1b[31m'* ]] || fail "null-login author diagnostic did not escape control characters"
[[ "$null_output" != *$'\033'* ]] || fail "null-login author diagnostic emitted a raw terminal escape"
[[ "$(printf '%s\n' "$null_output" | wc -l)" -eq 1 ]] || fail "null-login author diagnostic permitted newline injection"
[[ "$null_output" == *"tl-mosaic"* ]] || fail "null-login author omitted the named escalation principal"
[[ ! -e "$null_dir/merge-payload.json" ]] || fail "null-login BLOCK still reached the merge API"
# Every provider-derived field in alternate BLOCK diagnostics is log-safe too,
# including an invalid non-head SHA that contains control characters.
unsafe_dir=$(make_case unsafe-identity)
set +e
unsafe_output=$(run_case "$unsafe_dir" unsafe-identity --co-author-trailers --escalate-to tl-mosaic 2>&1)
unsafe_rc=$?
set -e
[[ "$unsafe_rc" -ne 0 ]] || fail "unsafe identity expected a non-zero BLOCK"
[[ "$unsafe_output" == *"unusable linked identity"* ]] || fail "unsafe identity omitted its BLOCK reason"
[[ "$unsafe_output" == *'\n\x1b[31m'* ]] || fail "unsafe identity SHA did not escape control characters"
[[ "$unsafe_output" != *$'\033'* ]] || fail "unsafe identity diagnostic emitted a raw terminal escape"
[[ "$(printf '%s\n' "$unsafe_output" | wc -l)" -eq 1 ]] || fail "unsafe identity diagnostic permitted newline injection"
[[ ! -e "$unsafe_dir/merge-payload.json" ]] || fail "unsafe identity BLOCK still reached the merge API"
# The provider PR title cannot add an unchecked trailer outside the constructed
# message field: multi-line and trailer-shaped titles block before mutation.
title_dir=$(make_case title-injection)
set +e
title_output=$(MOSAIC_TEST_TITLE_MODE=injection \
run_case "$title_dir" verified --co-author-trailers --escalate-to tl-mosaic 2>&1)
title_rc=$?
set -e
[[ "$title_rc" -ne 0 ]] || fail "title trailer injection unexpectedly passed"
[[ "$title_output" == *"not one printable, non-trailer line"* ]] || fail "title injection refusal lost its diagnostic"
[[ ! -e "$title_dir/merge-payload.json" ]] || fail "title injection reached the merge API"
# Provider failures remain diagnosable after their temporary response file is
# removed, but provider-controlled control characters stay log-safe.
error_dir=$(make_case provider-error)
set +e
error_output=$(MOSAIC_TEST_BASIC_AVAILABLE=true MOSAIC_TEST_FALLBACK_MODE=provider-error \
run_case "$error_dir" single 2>&1)
error_rc=$?
set -e
[[ "$error_rc" -ne 0 ]] || fail "provider error unexpectedly passed"
[[ "$error_output" == *"HTTP 409"* ]] || fail "provider error omitted the HTTP status"
[[ "$error_output" == *"branch policy rejected"* ]] || fail "provider error response was discarded"
[[ "$error_output" == *'\n\x1b[31m'* ]] || fail "provider error response did not escape control characters"
[[ "$error_output" != *$'\033'* ]] || fail "provider error response emitted a raw terminal escape"
[[ "$error_output" != *"Basic Auth fallback"* ]] || fail "provider error advertised removed Basic Auth fallback"
[[ ! -e "$error_dir/basic-resolution.log" ]] || fail "HTTP 409 policy denial incorrectly triggered Basic Auth fallback"
# Authorization denials likewise fail closed instead of changing principals.
forbidden_dir=$(make_case forbidden)
set +e
forbidden_output=$(MOSAIC_TEST_BASIC_AVAILABLE=true MOSAIC_TEST_FALLBACK_MODE=forbidden \
run_case "$forbidden_dir" single 2>&1)
forbidden_rc=$?
set -e
[[ "$forbidden_rc" -ne 0 ]] || fail "HTTP 403 authorization denial unexpectedly passed"
[[ "$forbidden_output" == *"HTTP 403"* ]] || fail "authorization denial omitted the HTTP status"
[[ "$forbidden_output" != *"Basic Auth fallback"* ]] || fail "authorization denial advertised removed Basic Auth fallback"
[[ ! -e "$forbidden_dir/basic-resolution.log" ]] || fail "HTTP 403 authorization denial incorrectly triggered Basic Auth fallback"
# The BLOCK destination cannot be generic or inferred after failure: opting in
# without a named principal is refused before any provider operation.
principal_dir=$(make_case missing-principal)
set +e
principal_output=$(run_case "$principal_dir" verified --co-author-trailers 2>&1)
principal_rc=$?
set -e
[[ "$principal_rc" -ne 0 ]] || fail "co-author mode without a named principal unexpectedly passed"
[[ "$principal_output" == *"requires --escalate-to with a named principal"* ]] || fail "missing-principal refusal lost its diagnostic"
[[ ! -e "$principal_dir/merge-payload.json" ]] || fail "missing-principal refusal reached the merge API"
# A trailing value-taking option receives a stable CLI diagnostic instead of a
# set -u unbound-variable crash.
value_dir=$(make_case missing-principal-value)
set +e
value_output=$(run_case "$value_dir" verified --co-author-trailers --escalate-to 2>&1)
value_rc=$?
set -e
[[ "$value_rc" -ne 0 ]] || fail "missing --escalate-to value unexpectedly passed"
[[ "$value_output" == *"--escalate-to requires one principal name"* ]] || fail "missing --escalate-to value lost its diagnostic"
[[ "$value_output" != *"unbound variable"* ]] || fail "missing --escalate-to value crashed under set -u"
[[ ! -e "$value_dir/merge-payload.json" ]] || fail "missing --escalate-to value reached the merge API"
# Negative control: ordinary single-author merge remains byte-for-byte payload
# compatible and hardcoded to squash, with no optional message fields.
single_dir=$(make_case single)
set +e
single_output=$(run_case "$single_dir" single 2>&1)
single_rc=$?
set -e
if [[ "$single_rc" -ne 0 ]]; then
fail "ordinary single-author merge expected rc=0, got rc=$single_rc: $single_output"
elif [[ ! -s "$single_dir/merge-payload.json" ]]; then
fail "ordinary single-author merge did not reach the API payload"
else
python3 - "$single_dir/merge-payload.json" <<'PY' || fail "ordinary single-author payload changed"
import json
import sys
payload = json.load(open(sys.argv[1], encoding="utf-8"))
assert payload == {
"Do": "squash",
"head_commit_id": "1111111111111111111111111111111111111111",
}, payload
PY
fi
[[ -e "$single_dir/auth-via-config" ]] || fail "ordinary path did not authenticate curl through stdin config"
[[ ! -e "$single_dir/token-in-argv" ]] || fail "ordinary path placed the Gitea token in curl argv"
[[ "$(wc -l < "$single_dir/token-resolution.log")" -eq 1 ]] || fail "ordinary path did not use exactly one credential resolution"
# Squash is not defaultable: an explicit non-squash method must remain refused.
method_dir=$(make_case method-refusal)
set +e
method_output=$(run_case "$method_dir" single -m merge 2>&1)
method_rc=$?
set -e
[[ "$method_rc" -ne 0 ]] || fail "non-squash method unexpectedly passed"
[[ "$method_output" == *"enforces squash merge only"* ]] || fail "non-squash refusal lost its policy diagnostic"
[[ ! -e "$method_dir/merge-payload.json" ]] || fail "non-squash refusal reached the merge API"
if [[ "$failures" -ne 0 ]]; then
echo "pr-merge message-field regression failed ($failures assertions)" >&2
exit 1
fi
echo "pr-merge message-field regression passed (verified, BLOCK, and unchanged squash control)"
@@ -39,7 +39,7 @@ MAX_FRAME: Final = 64 * 1024
MAX_STATE: Final = 4 * 1024 * 1024 MAX_STATE: Final = 4 * 1024 * 1024
MAX_PENDING_TOKENS: Final = 256 MAX_PENDING_TOKENS: Final = 256
MAX_IN_FLIGHT_CONNECTIONS: Final = 16 MAX_IN_FLIGHT_CONNECTIONS: Final = 16
MAX_LEASE_TTL_SECONDS: Final = 300 MAX_LEASE_TTL_SECONDS: Final = 3600
STATE_VERSION: Final = 1 STATE_VERSION: Final = 1
READ_DEADLINE_SECONDS: Final = 1.0 READ_DEADLINE_SECONDS: Final = 1.0
HANDLE_QUEUE_TIMEOUT_SECONDS: Final = 1.0 HANDLE_QUEUE_TIMEOUT_SECONDS: Final = 1.0
@@ -50,8 +50,8 @@ LEASE_PENDING: Final = "PENDING_VERIFICATION"
LEASE_PENDING_PROMOTION: Final = "PENDING_PROMOTION" LEASE_PENDING_PROMOTION: Final = "PENDING_PROMOTION"
LEASE_VERIFIED: Final = "VERIFIED" LEASE_VERIFIED: Final = "VERIFIED"
READ_ONLY_TOOLS: Final = { READ_ONLY_TOOLS: Final = {
"claude": frozenset({"Read", "Grep", "Glob", "Ls", "Find"}), "claude": frozenset({"Read", "Grep", "Glob"}),
"pi": frozenset({"read", "grep", "find", "ls"}), "pi": frozenset({"read", "ls"}),
} }
RECOVERY_TOOL: Final = "mosaic_context_recover" RECOVERY_TOOL: Final = "mosaic_context_recover"
@@ -8,7 +8,9 @@ import json
import os import os
import socket import socket
import sys import sys
import time
from collections.abc import Callable, Mapping, Sequence from collections.abc import Callable, Mapping, Sequence
from datetime import datetime, timezone
from pathlib import Path from pathlib import Path
from typing import Final from typing import Final
@@ -53,6 +55,48 @@ def broker_request(socket_path: Path, request: dict[str, object]) -> dict[str, o
return value return value
def _self_starttime() -> str | None:
"""Field 22 of our own /proc stat — the anchor starttime the broker records.
Read past the comm field's parens, since a process name may contain them.
"""
try:
raw = Path(f"/proc/{os.getpid()}/stat").read_text()
return raw.rsplit(")", 1)[1].split()[19]
except (OSError, IndexError, ValueError):
return None
def _append_launch_record(environ: Mapping[str, str], record: dict[str, object]) -> None:
"""Append one NDJSON event to the #797 Runtime Session Ledger.
`fleet/run/sessions/` is operator-classified in framework-manifest.txt and is
already covered by test-upgrade-manifest-guard.sh, so an upgrade can neither
overwrite nor prune it. Files 0600 under a 0700 dir, matching what that guard
asserts.
Never raises: a launch must not be denied over bookkeeping. But it also never
fails silently a missing record is exactly the kind of gap that made the
2026-08-06 MUTATOR_UNVERIFIED investigation cost a day.
"""
try:
mosaic_home = environ.get("MOSAIC_HOME") or str(Path.home() / ".config" / "mosaic")
directory = Path(mosaic_home) / "fleet" / "run" / "sessions"
directory.mkdir(parents=True, exist_ok=True)
os.chmod(directory, 0o700)
framed = {
"seq": time.time_ns() // 1_000_000,
"ts": datetime.now(timezone.utc).isoformat(),
**record,
}
path = directory / "events.ndjson"
descriptor = os.open(path, os.O_WRONLY | os.O_CREAT | os.O_APPEND, 0o600)
with os.fdopen(descriptor, "w") as handle:
handle.write(json.dumps(framed, separators=(",", ":")) + "\n")
except (OSError, ValueError, TypeError) as error:
print(f"[mosaic] WARNING: launch record not written: {error}", file=sys.stderr)
def main( def main(
argv: Sequence[str] | None = None, argv: Sequence[str] | None = None,
*, *,
@@ -94,8 +138,9 @@ def main(
# silent pass and never folded into the generic registration-failure # silent pass and never folded into the generic registration-failure
# branch. # branch.
try: try:
activation_capability = probe_activation_capability(source_environment)
assert_activation_capability_matches( assert_activation_capability_matches(
probe_activation_capability(source_environment), activation_capability,
expected_activation_capability, expected_activation_capability,
) )
except VersionCouplingError as version_error: except VersionCouplingError as version_error:
@@ -128,6 +173,32 @@ def main(
print("Mosaic lease broker registration failed; runtime launch denied.", file=sys.stderr) print("Mosaic lease broker registration failed; runtime launch denied.", file=sys.stderr)
return 1 return 1
# Immutable launch record, half two. `mosaic` wrote `session.launch` with the
# config/provenance it knows; only this process knows the broker session id
# and the activation capability it just asserted. os.execvpe preserves the
# PID, so this PID is BOTH the anchor pid and the join key back to that
# record. Never fatal — bookkeeping must not deny a launch — but never
# silent either.
_append_launch_record(
source_environment,
{
"kind": "lease.register",
# Joins back to `mosaic`'s session.launch record. NOT pid: execRuntime()
# spawns rather than execs, so this process is a CHILD of mosaic with a
# different pid. This pid IS the broker anchor pid (os.execvpe below
# preserves it), which is a separate and still-useful fact.
"launch_id": source_environment.get("MOSAIC_LAUNCH_ID"),
"pid": os.getpid(),
"runtime": arguments.runtime,
"session_id": session_id,
"runtime_generation": generation,
"generation_file": str(generation_file),
"anchor_starttime": _self_starttime(),
"activation_capability": activation_capability,
"command": Path(command[0]).name,
},
)
environment = dict(source_environment) environment = dict(source_environment)
environment["MOSAIC_LEASE_SESSION_ID"] = session_id environment["MOSAIC_LEASE_SESSION_ID"] = session_id
environment["MOSAIC_RUNTIME_GENERATION"] = str(generation) environment["MOSAIC_RUNTIME_GENERATION"] = str(generation)
@@ -0,0 +1,337 @@
#!/usr/bin/env python3
"""Lease promotion client — the half the enforcement toolkit never shipped.
The enforcement half (``daemon.py`` + ``mutator-gate.py``) ships and denies. The
promotion half has no production caller anywhere in the package: as of 0.0.48,
0.0.49 and 0.0.50-next.2207, ``begin_verification`` / ``observe_receipt`` /
``promote_lease`` are invoked only by ``broker-test-client.ts``, the acceptance
spec, unit tests, and two probes under ``docs/``. Consequence: **no lease on any
host can reach VERIFIED**, so every mutator is denied ``MUTATOR_UNVERIFIED`` by a
gate nothing can satisfy.
THE PROTOCOL (``daemon.py:578-754``)
------------------------------------
1. ``begin_verification`` broker revokes, mints a challenge, and returns the
exact ``receipt`` text the MODEL must emit
2. *the model emits that text verbatim as its ENTIRE latest message*
3. the runtime adapter ships that message to the daemon-owned observer socket
4. ``observe_receipt`` -> ``PENDING_PROMOTION``
5. ``promote_lease`` -> ``VERIFIED``
THIS MODULE IMPLEMENTS 1, 4 AND 5 NEVER 2
-------------------------------------------
Step 2 is the security property, not a formality. ``is_verbatim_receipt`` uses
``hmac.compare_digest`` against the exact minted string explicitly "not a
transcript substring" (``receipt_challenge.py``). Promotion therefore requires a
live model that received the challenge in its context and echoed it exactly.
``receipt-observer-client.py`` will post ANY string as the latest assistant
message. A promotion client that posted its own receipt would satisfy the broker
while proving nothing a gate-disabler indistinguishable from a working fix
unless someone looks for it. **This module never posts a receipt.** Emitting it
belongs to the runtime adapter, where a real model turn happens.
The construction binds the exact normative source bytes. ``h_source`` /
``h_payload`` are derived by the framework's own
``normative_fragments.build_payload`` rather than reimplemented: the broker
derives them the same way and any divergence yields ``PAYLOAD_BINDING_MISMATCH``.
There must be exactly one implementation.
WHAT THE BINDING DOES *NOT* PROVE
---------------------------------
It is tempting to read a VERIFIED lease as "this agent is running THIS law".
**It does not mean that**, and writing it down that way is how the belief spread.
The broker holds no reference copy of any normative source and never opens one;
it recomputes ``h_source`` / ``h_payload`` from the fragment bytes THIS CLIENT
sent and compares them to the binding THIS CLIENT sent (``daemon.py:602-616``).
Both sides of that comparison originate here, so it detects corruption in
transit and nothing else. What the binding actually asserts is "the client
claims these bytes, self-consistently".
Making it mean the stronger thing requires the broker to re-read the on-disk
sources itself, against a manifest the agent cannot rewrite i.e. broker code
attestation under its own uid. Until then, do not cite a VERIFIED lease as
evidence of law integrity.
Usage
-----
lease_promote.py --begin # prints the receipt the MODEL must emit
lease_promote.py --complete <challenge> # after the adapter observed it
"""
from __future__ import annotations
import argparse
import base64
import hashlib
import json
import os
import socket
import sys
from pathlib import Path
from typing import Final
# Isolated (`python -I`) adapter invocations must still import co-located
# framework modules; never depend on the caller's PYTHONPATH.
_MODULE_DIRECTORY = str(Path(__file__).resolve().parent)
if _MODULE_DIRECTORY not in sys.path:
sys.path.insert(0, _MODULE_DIRECTORY)
from normative_fragments import NormativeFragment, build_payload # noqa: E402
MAX_FRAME: Final = 64 * 1024
BROKER_TIMEOUT_SECONDS: Final = 3.0
SCHEMA_VERSION: Final = 1
MANIFEST_VERSION: Final = 1
GENERATOR_VERSION: Final = "mosaic/lease_promote@1"
DEFAULT_TTL_SECONDS: Final = 3600
# Normative sources whose exact bytes bind the lease, in binding order. Order is
# load-bearing: ``h_source`` frames the resolved sequence, so reordering changes
# the derivation. Never fabricate a source that is not on disk.
FRAGMENT_SOURCES: Final = (
"CONSTITUTION.md",
"AGENTS.md",
"SOUL.md",
"USER.md",
"STANDARDS.md",
"TOOLS.md",
)
# Framework-owned sources, reconciled on every upgrade — `install.sh:76`
# FRAMEWORK_OWNED and `config/file-adapter.ts` FRAMEWORK_OWNED_FILES — plus the
# per-runtime contract shipped under `framework/runtime/<runtime>/`. A deployment
# missing one of these is broken, not minimal, so their absence is refused rather
# than silently dropped from the binding.
#
# SOUL.md and USER.md are deliberately excluded: install.sh does not seed them
# ("intentionally NOT seeded here — they are generated by `mosaic init`"), so a
# fresh install legitimately lacks both. TOOLS.md is user-seeded on first install
# only. Absence of those three is reported, not fatal.
REQUIRED_SOURCES: Final = frozenset({"CONSTITUTION.md", "AGENTS.md", "STANDARDS.md"})
class IncompleteBinding(RuntimeError):
"""A source that must bind this lease could not be read.
**Never downgrade this to a skip.** The broker recomputes the hashes from the
fragments it is sent, so an omitted fragment is internally consistent and
``PAYLOAD_BINDING_MISMATCH`` cannot fire a partial law promotes exactly like
a complete one, and nothing downstream can tell the difference. Dropping an
unreadable source therefore does not degrade the binding, it forges a smaller
one. Fail here, where the omission is still visible.
"""
def mosaic_home() -> Path:
return Path(os.environ.get("MOSAIC_HOME") or Path.home() / ".config" / "mosaic")
def broker_socket() -> Path:
value = os.environ.get("MOSAIC_LEASE_BROKER_SOCKET")
if value:
return Path(value)
runtime_dir = os.environ.get("XDG_RUNTIME_DIR")
if runtime_dir:
return Path(runtime_dir) / "mosaic-lease" / "broker.sock"
return Path(f"/run/user/{os.getuid()}/mosaic-lease/broker.sock")
def session_identity() -> tuple[str, int, str]:
"""Session id, CURRENT generation, runtime.
The generation file wins over the env var, matching ``lease_generation.py``.
Sending a generation HIGHER than the broker's would revoke this session's own
authority (``daemon.py:342-344``), so this never guesses.
"""
session_id = os.environ["MOSAIC_LEASE_SESSION_ID"]
runtime = os.environ["MOSAIC_LEASE_RUNTIME"]
state_file = os.environ.get("MOSAIC_LEASE_GENERATION_FILE")
if state_file:
try:
return session_id, int(Path(state_file).read_text().strip()), runtime
except (OSError, ValueError):
pass
return session_id, int(os.environ["MOSAIC_RUNTIME_GENERATION"]), runtime
def build_construction(runtime: str) -> tuple[dict[str, object], object]:
"""Assemble the wire construction and derive its hashes with the sole builder."""
runtime_contract = f"runtime/{runtime}/RUNTIME.md"
sources = list(FRAGMENT_SOURCES) + [runtime_contract]
required = REQUIRED_SOURCES | {runtime_contract}
wire_fragments: list[dict[str, str]] = []
objects: list[NormativeFragment] = []
absent: list[str] = []
for source_id in sources:
try:
content = (mosaic_home() / source_id).read_bytes()
except FileNotFoundError:
# Genuinely not on disk. Legitimate only for operator-owned sources.
if source_id in required:
raise IncompleteBinding(
f"required normative source is absent: {source_id}"
) from None
absent.append(source_id)
continue
except OSError as exc:
# The path resolves but will not read — EACCES, EIO, EISDIR, ELOOP.
# That is an anomaly for EVERY source, optional ones included: an
# unreadable file is not an un-configured one, and treating it as
# absent is what lets a permission change quietly shrink the law.
raise IncompleteBinding(
f"normative source is present but unreadable: {source_id} "
f"({type(exc).__name__})"
) from exc
digest = hashlib.sha256(content).hexdigest()
wire_fragments.append(
{
"source_id": source_id,
"content_base64": base64.b64encode(content).decode("ascii"),
"expected_sha256": digest,
}
)
objects.append(NormativeFragment(source_id, content, digest))
if not wire_fragments:
raise IncompleteBinding("no normative sources found — refusing an empty binding")
# Absence is legitimate here but never invisible. The omission is already
# baked into h_source (the framed source sequence differs), but nothing
# compares h_source to an expected value, so this line is the only place a
# human learns the binding was narrower than the full set.
if absent:
print(
f"lease_promote: binding omits absent operator sources: {', '.join(absent)}",
file=sys.stderr,
)
result = build_payload(
manifest_version=MANIFEST_VERSION,
generator_version=GENERATOR_VERSION,
fragments=objects,
)
if result.injectionDecision != "ACCEPTED" or not result.promotion:
raise RuntimeError(f"construction refused locally: {result.source_reason}")
return (
{
"manifest_version": MANIFEST_VERSION,
"generator_version": GENERATOR_VERSION,
"fragments": wire_fragments,
},
result,
)
def broker_request(payload: dict[str, object]) -> dict[str, object]:
raw = (json.dumps(payload, separators=(",", ":")) + "\n").encode()
if len(raw) > MAX_FRAME:
raise ValueError(
f"request too large ({len(raw)} bytes); broker frame cap is {MAX_FRAME}"
)
response = bytearray()
with socket.socket(socket.AF_UNIX, socket.SOCK_STREAM) as connection:
connection.settimeout(BROKER_TIMEOUT_SECONDS)
connection.connect(str(broker_socket()))
connection.sendall(raw)
connection.shutdown(socket.SHUT_WR)
while len(response) <= MAX_FRAME:
chunk = connection.recv(4096)
if not chunk:
break
response.extend(chunk)
if len(response) > MAX_FRAME or not response.endswith(b"\n"):
raise ValueError("invalid broker reply")
value = json.loads(response)
if not isinstance(value, dict):
raise ValueError("invalid broker reply")
return value
def begin(
ttl_seconds: int = DEFAULT_TTL_SECONDS,
compaction_epoch: int = 0,
request_epoch: int = 0,
) -> dict[str, object]:
"""Step 1. Returns the broker reply, including the exact ``receipt`` text."""
session_id, generation, runtime = session_identity()
construction, derived = build_construction(runtime)
return broker_request(
{
"action": "begin_verification",
"session_id": session_id,
"runtime_generation": generation,
"runtime": runtime,
"ttl_seconds": ttl_seconds,
"binding": {
"compaction_epoch": compaction_epoch,
"request_epoch": request_epoch,
"h_source": derived.h_source,
"h_payload": derived.h_payload,
"schema_version": SCHEMA_VERSION,
},
"construction": construction,
}
)
def complete(challenge: str) -> dict[str, object]:
"""Steps 4-5. Assumes the model already emitted the receipt and the adapter
shipped it to the observer socket."""
session_id, generation, _ = session_identity()
observed = broker_request(
{
"action": "observe_receipt",
"session_id": session_id,
"runtime_generation": generation,
"receipt_challenge": challenge,
}
)
if observed.get("ok") is not True or observed.get("state") != "PENDING_PROMOTION":
return {"stage": "observe_receipt", **observed}
promoted = broker_request(
{
"action": "promote_lease",
"session_id": session_id,
"runtime_generation": generation,
"receipt_challenge": challenge,
}
)
return {"stage": "promote_lease", **promoted}
def main(argv: list[str] | None = None) -> int:
parser = argparse.ArgumentParser(description="Mosaic lease promotion client.")
group = parser.add_mutually_exclusive_group(required=True)
group.add_argument(
"--begin",
action="store_true",
help="mint a challenge; prints the receipt the MODEL must emit verbatim",
)
group.add_argument(
"--complete",
metavar="CHALLENGE",
help="observe the emitted receipt and promote the lease",
)
parser.add_argument("--ttl-seconds", type=int, default=DEFAULT_TTL_SECONDS)
arguments = parser.parse_args(argv)
try:
if arguments.begin:
print(json.dumps(begin(ttl_seconds=arguments.ttl_seconds), indent=2))
else:
print(json.dumps(complete(arguments.complete), indent=2))
except KeyError as exc:
print(f"missing lease environment: {exc}; not a lease-gated session", file=sys.stderr)
return 2
except (OSError, ValueError, RuntimeError, json.JSONDecodeError) as exc:
print(f"{type(exc).__name__}: {exc}", file=sys.stderr)
return 2
return 0
if __name__ == "__main__":
raise SystemExit(main())
@@ -0,0 +1,399 @@
#!/usr/bin/env python3
"""Claude UserPromptSubmit hook for operator-triggered lease promotion."""
from __future__ import annotations
import fcntl
import importlib.util
import json
import os
import secrets
import stat
import subprocess
import sys
import time
from collections.abc import Callable, Mapping
from pathlib import Path
from typing import Final, TextIO
_MODULE_DIRECTORY = str(Path(__file__).resolve().parent)
if _MODULE_DIRECTORY not in sys.path:
sys.path.insert(0, _MODULE_DIRECTORY)
from receipt_challenge import receipt_for # noqa: E402
_observer_spec = importlib.util.spec_from_file_location(
"mosaic_receipt_observer_client", Path(__file__).resolve().with_name("receipt-observer-client.py")
)
if _observer_spec is None or _observer_spec.loader is None:
raise RuntimeError("unable to load receipt observer client")
_observer_module = importlib.util.module_from_spec(_observer_spec)
_observer_spec.loader.exec_module(_observer_module)
observer_request = _observer_module.observer_request
MAX_FRAME: Final = 64 * 1024
PENDING_MAX_AGE_SECONDS: Final = 60 * 60
PROMOTER_TIMEOUT_SECONDS: Final = 10.0
PROMOTION_PROMPT: Final = "/mosaic-promote"
PROMOTER: Final = Path(__file__).resolve().with_name("lease_promote.py")
PENDING_DIRECTORY: Final = "mosaic-lease"
AUTHORIZATION_DIRECTORY: Final = "authorizations"
AUTHORIZATION_TTL_SECONDS: Final = 60
LEASE_TTL_SECONDS: Final = 60 * 60
LOCK_FILE: Final = "promotion.lock"
RESULT_FILE: Final = "last-result.json"
EXPECTED_BEGIN_KEYS: Final = frozenset(
{"ok", "state", "receipt_challenge", "receipt", "binding"}
)
EXPECTED_BINDING_KEYS: Final = frozenset(
{
"compaction_epoch",
"request_epoch",
"h_source",
"h_payload",
"runtime_generation",
"schema_version",
}
)
class PromotionAlreadyInProgress(RuntimeError):
pass
def reject_duplicate_json_keys(pairs: list[tuple[str, object]]) -> dict[str, object]:
value: dict[str, object] = {}
for key, item in pairs:
if key in value:
raise ValueError("duplicate promoter JSON key")
value[key] = item
return value
def read_hook_input(stream: object) -> dict[str, object]:
raw = getattr(stream, "buffer", stream).read(MAX_FRAME + 1)
if not isinstance(raw, bytes) or len(raw) > MAX_FRAME:
raise ValueError("invalid UserPromptSubmit input")
value = json.loads(raw, object_pairs_hook=reject_duplicate_json_keys)
if not isinstance(value, dict):
raise ValueError("invalid UserPromptSubmit input")
return value
def emit_context(stream: TextIO, message: str) -> None:
json.dump(
{
"hookSpecificOutput": {
"hookEventName": "UserPromptSubmit",
"additionalContext": message,
}
},
stream,
separators=(",", ":"),
)
stream.write("\n")
def session_pending_name(environ: Mapping[str, str]) -> tuple[Path, str]:
runtime_dir = Path(environ["XDG_RUNTIME_DIR"])
session_id = environ["MOSAIC_LEASE_SESSION_ID"]
if not runtime_dir.is_absolute():
raise ValueError("XDG_RUNTIME_DIR must be absolute")
if len(session_id) != 64 or any(character not in "0123456789abcdef" for character in session_id):
raise ValueError("invalid lease session id")
return runtime_dir, f"pending-{session_id}"
def open_pending_directory(runtime_dir: Path) -> int:
directory_flags = (
os.O_RDONLY
| getattr(os, "O_CLOEXEC", 0)
| getattr(os, "O_DIRECTORY", 0)
| getattr(os, "O_NOFOLLOW", 0)
)
runtime_descriptor = os.open(runtime_dir, directory_flags)
try:
runtime_metadata = os.fstat(runtime_descriptor)
if (
not stat.S_ISDIR(runtime_metadata.st_mode)
or runtime_metadata.st_uid != os.getuid()
or stat.S_IMODE(runtime_metadata.st_mode) != 0o700
):
raise ValueError("unsafe XDG runtime directory")
try:
os.mkdir(PENDING_DIRECTORY, mode=0o700, dir_fd=runtime_descriptor)
except FileExistsError:
pass
descriptor = os.open(PENDING_DIRECTORY, directory_flags, dir_fd=runtime_descriptor)
finally:
os.close(runtime_descriptor)
metadata = os.fstat(descriptor)
if (
not stat.S_ISDIR(metadata.st_mode)
or metadata.st_uid != os.getuid()
or stat.S_IMODE(metadata.st_mode) != 0o700
):
os.close(descriptor)
raise ValueError("unsafe promotion pending directory")
return descriptor
def acquire_lock(directory_descriptor: int) -> int:
flags = (
os.O_RDWR
| os.O_CREAT
| getattr(os, "O_CLOEXEC", 0)
| getattr(os, "O_NOFOLLOW", 0)
)
descriptor = os.open(LOCK_FILE, flags, 0o600, dir_fd=directory_descriptor)
metadata = os.fstat(descriptor)
if (
not stat.S_ISREG(metadata.st_mode)
or metadata.st_uid != os.getuid()
or stat.S_IMODE(metadata.st_mode) != 0o600
):
os.close(descriptor)
raise ValueError("unsafe promotion lock file")
try:
fcntl.flock(descriptor, fcntl.LOCK_EX | fcntl.LOCK_NB)
except BlockingIOError as error:
os.close(descriptor)
raise PromotionAlreadyInProgress() from error
return descriptor
def sweep_stale_pending(directory_descriptor: int, current_time: float) -> None:
cutoff = current_time - PENDING_MAX_AGE_SECONDS
removed = False
with os.scandir(directory_descriptor) as entries:
for candidate in entries:
if not (
candidate.name.startswith("pending-")
or candidate.name.startswith(".pending-")
):
continue
try:
metadata = candidate.stat(follow_symlinks=False)
if metadata.st_mtime < cutoff and not stat.S_ISDIR(metadata.st_mode):
os.unlink(candidate.name, dir_fd=directory_descriptor)
removed = True
except FileNotFoundError:
continue
if removed:
os.fsync(directory_descriptor)
def consume_authorization(directory_descriptor: int, session_id: str, wall_clock: float) -> str | None:
flags = os.O_RDONLY | getattr(os, "O_CLOEXEC", 0) | getattr(os, "O_DIRECTORY", 0) | getattr(os, "O_NOFOLLOW", 0)
try:
authorization_descriptor = os.open(AUTHORIZATION_DIRECTORY, flags, dir_fd=directory_descriptor)
except FileNotFoundError:
return None
try:
metadata = os.fstat(authorization_descriptor)
if not stat.S_ISDIR(metadata.st_mode) or metadata.st_uid != os.getuid() or stat.S_IMODE(metadata.st_mode) != 0o700:
raise ValueError("unsafe promotion authorization directory")
name = f"{session_id}.auth"
try:
descriptor = os.open(name, os.O_RDONLY | getattr(os, "O_CLOEXEC", 0) | getattr(os, "O_NOFOLLOW", 0), dir_fd=authorization_descriptor)
except FileNotFoundError:
return None
try:
token_metadata = os.fstat(descriptor)
if not stat.S_ISREG(token_metadata.st_mode) or token_metadata.st_uid != os.getuid() or stat.S_IMODE(token_metadata.st_mode) != 0o600 or token_metadata.st_size <= 0 or token_metadata.st_size > MAX_FRAME:
raise ValueError("unsafe promotion authorization")
raw = os.read(descriptor, MAX_FRAME + 1)
finally:
os.close(descriptor)
os.unlink(name, dir_fd=authorization_descriptor)
os.fsync(authorization_descriptor)
token = json.loads(raw, object_pairs_hook=reject_duplicate_json_keys)
if not isinstance(token, dict) or set(token) != {"nonce", "seat", "session_id", "expires_at", "ts"}:
return None
nonce = token.get("nonce")
expires_at = token.get("expires_at")
issued_at = token.get("ts")
if token.get("session_id") != session_id or not isinstance(token.get("seat"), str) or not isinstance(nonce, str) or len(nonce) != 64 or any(char not in "0123456789abcdef" for char in nonce) or type(expires_at) not in (int, float) or type(issued_at) not in (int, float) or expires_at <= wall_clock or expires_at > issued_at + AUTHORIZATION_TTL_SECONDS:
return None
return nonce
finally:
os.close(authorization_descriptor)
def write_result(directory_descriptor: int, attempt_id: str, verified: bool, reason: str | None, session_id: str, wall_clock: float) -> None:
temporary = f".{RESULT_FILE}.tmp-{secrets.token_hex(8)}"
descriptor = os.open(temporary, os.O_WRONLY | os.O_CREAT | os.O_EXCL | getattr(os, "O_CLOEXEC", 0) | getattr(os, "O_NOFOLLOW", 0), 0o600, dir_fd=directory_descriptor)
try:
os.fchmod(descriptor, 0o600)
with os.fdopen(descriptor, "w", encoding="utf-8", closefd=False) as stream:
json.dump({"attempt_id": attempt_id, "expires_at_wallclock": wall_clock + LEASE_TTL_SECONDS if verified else None, "reason": reason, "session_id": session_id, "ts": wall_clock, "verified": verified}, stream, separators=(",", ":"), sort_keys=True)
stream.flush(); os.fsync(stream.fileno())
os.replace(temporary, RESULT_FILE, src_dir_fd=directory_descriptor, dst_dir_fd=directory_descriptor)
os.fsync(directory_descriptor)
finally:
os.close(descriptor)
def write_pending(directory_descriptor: int, name: str, challenge: str) -> None:
temporary = f".{name}.tmp-{secrets.token_hex(8)}"
flags = (
os.O_WRONLY
| os.O_CREAT
| os.O_EXCL
| getattr(os, "O_CLOEXEC", 0)
| getattr(os, "O_NOFOLLOW", 0)
)
descriptor = os.open(temporary, flags, 0o600, dir_fd=directory_descriptor)
try:
os.fchmod(descriptor, 0o600)
with os.fdopen(descriptor, "w", encoding="utf-8", closefd=False) as stream:
stream.write(challenge)
stream.flush()
os.fsync(stream.fileno())
os.replace(
temporary,
name,
src_dir_fd=directory_descriptor,
dst_dir_fd=directory_descriptor,
)
os.fsync(directory_descriptor)
except Exception:
try:
os.unlink(temporary, dir_fd=directory_descriptor)
except FileNotFoundError:
pass
raise
finally:
os.close(descriptor)
def parse_begin_reply(
completed: subprocess.CompletedProcess[str],
) -> tuple[str, dict[str, object] | None]:
if completed.returncode != 0:
return f"PROMOTER_EXIT_{completed.returncode}", None
try:
value = json.loads(
completed.stdout,
object_pairs_hook=reject_duplicate_json_keys,
)
except (json.JSONDecodeError, RecursionError, TypeError, ValueError):
return "INVALID_PROMOTER_REPLY", None
if not isinstance(value, dict):
return "INVALID_PROMOTER_REPLY", None
if value.get("ok") is False and set(value) == {"ok", "code"}:
code = value.get("code")
return code if isinstance(code, str) and code else "PROMOTION_BEGIN_REFUSED", value
if set(value) != EXPECTED_BEGIN_KEYS or value.get("ok") is not True:
return "INVALID_PROMOTER_REPLY", None
if value.get("state") != "PENDING_VERIFICATION":
return "INVALID_PROMOTER_REPLY", None
challenge = value.get("receipt_challenge")
receipt = value.get("receipt")
binding = value.get("binding")
if (
not isinstance(challenge, str)
or len(challenge) != 64
or any(character not in "0123456789abcdef" for character in challenge)
or not isinstance(receipt, str)
or not isinstance(binding, dict)
or set(binding) != EXPECTED_BINDING_KEYS
):
return "INVALID_PROMOTER_REPLY", None
integer_fields = (
"compaction_epoch",
"request_epoch",
"runtime_generation",
"schema_version",
)
if any(type(binding.get(field)) is not int or binding[field] < 0 for field in integer_fields):
return "INVALID_PROMOTER_REPLY", None
if not all(
isinstance(binding.get(field), str)
and len(binding[field]) == 64
and all(character in "0123456789abcdef" for character in binding[field])
for field in ("h_source", "h_payload")
):
return "INVALID_PROMOTER_REPLY", None
if not secrets.compare_digest(
receipt.encode("utf-8"),
receipt_for(challenge, binding).encode("utf-8"),
):
return "INVALID_PROMOTER_REPLY", None
return "", value
def main(
*,
environ: Mapping[str, str] | None = None,
stdin: object | None = None,
stdout: TextIO | None = None,
stderr: TextIO | None = None,
run: Callable[..., subprocess.CompletedProcess[str]] = subprocess.run,
now: Callable[[], float] = time.time,
) -> int:
source_environment = os.environ if environ is None else environ
input_stream = sys.stdin if stdin is None else stdin
output_stream = sys.stdout if stdout is None else stdout
error_stream = sys.stderr if stderr is None else stderr
try:
hook_input = read_hook_input(input_stream)
except (OSError, RecursionError, ValueError, json.JSONDecodeError) as error:
print(f"Mosaic promotion trigger ignored invalid hook input: {error}", file=error_stream)
return 0
if hook_input.get("prompt") != PROMOTION_PROMPT:
return 0
directory_descriptor: int | None = None
lock_descriptor: int | None = None
try:
runtime_dir, pending_name = session_pending_name(source_environment)
session_id = source_environment["MOSAIC_LEASE_SESSION_ID"]
directory_descriptor = open_pending_directory(runtime_dir)
lock_descriptor = acquire_lock(directory_descriptor)
wall_clock = now()
nonce = consume_authorization(directory_descriptor, session_id, wall_clock)
if nonce is None:
write_result(directory_descriptor, "0" * 64, False, "NOT_AUTHORIZED", session_id, wall_clock)
print("Mosaic promotion denied: NOT_AUTHORIZED.", file=error_stream)
return 0
sweep_stale_pending(directory_descriptor, wall_clock)
completed = run([sys.executable, "-I", "-S", "-B", str(PROMOTER), "--begin"], check=False, capture_output=True, text=True, env=dict(source_environment), timeout=PROMOTER_TIMEOUT_SECONDS)
code, reply = parse_begin_reply(completed)
if code or reply is None:
write_result(directory_descriptor, nonce, False, code or "PROMOTION_BEGIN_FAILED", session_id, now())
return 0
challenge = str(reply["receipt_challenge"])
observation = observer_request(
Path(source_environment["MOSAIC_RECEIPT_OBSERVER_SOCKET"]),
{"action": "record_runtime_observation", "session_id": session_id, "runtime_generation": int(source_environment["MOSAIC_RUNTIME_GENERATION"]), "runtime": "claude", "latest_assistant_message": reply["receipt"]},
)
if set(observation) != {"ok"} or observation.get("ok") is not True:
write_result(directory_descriptor, challenge, False, "OBSERVATION_REJECTED", session_id, now())
return 0
completion = run([sys.executable, "-I", "-S", "-B", str(PROMOTER), "--complete", challenge], check=False, capture_output=True, text=True, env=dict(source_environment), timeout=PROMOTER_TIMEOUT_SECONDS)
try:
outcome = json.loads(completion.stdout, object_pairs_hook=reject_duplicate_json_keys)
except (json.JSONDecodeError, ValueError):
outcome = None
if completion.returncode == 0 and isinstance(outcome, dict) and outcome.get("stage") == "promote_lease" and outcome.get("ok") is True and outcome.get("state") == "VERIFIED":
write_result(directory_descriptor, challenge, True, None, session_id, now())
else:
reason = outcome.get("code") if isinstance(outcome, dict) and isinstance(outcome.get("code"), str) else "PROMOTION_INCOMPLETE"
write_result(directory_descriptor, challenge, False, reason, session_id, now())
except PromotionAlreadyInProgress:
print("Mosaic promotion denied: PROMOTION_ALREADY_IN_PROGRESS.", file=error_stream)
except (KeyError, OSError, RecursionError, ValueError, subprocess.SubprocessError) as error:
print(f"Mosaic promotion begin failed: {type(error).__name__}: {error}", file=error_stream)
finally:
if lock_descriptor is not None:
os.close(lock_descriptor)
if directory_descriptor is not None:
os.close(directory_descriptor)
return 0
if __name__ == "__main__":
raise SystemExit(main())
@@ -0,0 +1,361 @@
#!/usr/bin/env python3
"""Claude Stop hook that completes a pending operator-triggered promotion."""
from __future__ import annotations
import fcntl
import json
import os
import secrets
import stat
import subprocess
import sys
import time
from collections.abc import Callable, Mapping
from pathlib import Path
from typing import Final, NamedTuple, TextIO
MAX_FRAME: Final = 64 * 1024
PROMOTER_TIMEOUT_SECONDS: Final = 10.0
LEASE_TTL_SECONDS: Final = 60 * 60
PROMOTER: Final = Path(__file__).resolve().with_name("lease_promote.py")
PENDING_DIRECTORY: Final = "mosaic-lease"
LOCK_FILE: Final = "promotion.lock"
RESULT_FILE: Final = "last-result.json"
TERMINAL_FAILURE_CODES: Final = frozenset(
{
"RECEIPT_REPLAY",
"RECEIPT_MISMATCH",
"INVALID_LEASE_TRANSITION",
"PROMOTION_TOKEN_INVALID",
}
)
class PendingChallenge(NamedTuple):
value: str
device: int
inode: int
def reject_duplicate_json_keys(pairs: list[tuple[str, object]]) -> dict[str, object]:
value: dict[str, object] = {}
for key, item in pairs:
if key in value:
raise ValueError("duplicate promoter JSON key")
value[key] = item
return value
def session_pending_name(environ: Mapping[str, str]) -> tuple[Path, str]:
runtime_dir = Path(environ["XDG_RUNTIME_DIR"])
session_id = environ["MOSAIC_LEASE_SESSION_ID"]
if not runtime_dir.is_absolute():
raise ValueError("XDG_RUNTIME_DIR must be absolute")
if len(session_id) != 64 or any(character not in "0123456789abcdef" for character in session_id):
raise ValueError("invalid lease session id")
return runtime_dir, f"pending-{session_id}"
def open_pending_directory(runtime_dir: Path) -> int | None:
directory_flags = (
os.O_RDONLY
| getattr(os, "O_CLOEXEC", 0)
| getattr(os, "O_DIRECTORY", 0)
| getattr(os, "O_NOFOLLOW", 0)
)
try:
runtime_descriptor = os.open(runtime_dir, directory_flags)
except FileNotFoundError:
return None
try:
runtime_metadata = os.fstat(runtime_descriptor)
if (
not stat.S_ISDIR(runtime_metadata.st_mode)
or runtime_metadata.st_uid != os.getuid()
or stat.S_IMODE(runtime_metadata.st_mode) != 0o700
):
raise ValueError("unsafe XDG runtime directory")
try:
descriptor = os.open(PENDING_DIRECTORY, directory_flags, dir_fd=runtime_descriptor)
except FileNotFoundError:
return None
finally:
os.close(runtime_descriptor)
metadata = os.fstat(descriptor)
if (
not stat.S_ISDIR(metadata.st_mode)
or metadata.st_uid != os.getuid()
or stat.S_IMODE(metadata.st_mode) != 0o700
):
os.close(descriptor)
raise ValueError("unsafe promotion pending directory")
return descriptor
def acquire_lock(directory_descriptor: int) -> int:
flags = (
os.O_RDWR
| os.O_CREAT
| getattr(os, "O_CLOEXEC", 0)
| getattr(os, "O_NOFOLLOW", 0)
)
descriptor = os.open(LOCK_FILE, flags, 0o600, dir_fd=directory_descriptor)
metadata = os.fstat(descriptor)
if (
not stat.S_ISREG(metadata.st_mode)
or metadata.st_uid != os.getuid()
or stat.S_IMODE(metadata.st_mode) != 0o600
):
os.close(descriptor)
raise ValueError("unsafe promotion lock file")
try:
fcntl.flock(descriptor, fcntl.LOCK_EX | fcntl.LOCK_NB)
except BlockingIOError:
os.close(descriptor)
raise
return descriptor
def read_pending(directory_descriptor: int, name: str) -> PendingChallenge | None:
flags = os.O_RDONLY | getattr(os, "O_CLOEXEC", 0) | getattr(os, "O_NOFOLLOW", 0)
try:
descriptor = os.open(name, flags, dir_fd=directory_descriptor)
except FileNotFoundError:
return None
try:
metadata = os.fstat(descriptor)
if (
not stat.S_ISREG(metadata.st_mode)
or metadata.st_uid != os.getuid()
or stat.S_IMODE(metadata.st_mode) != 0o600
or metadata.st_size <= 0
or metadata.st_size > MAX_FRAME
):
raise ValueError("unsafe promotion pending file")
raw = os.read(descriptor, MAX_FRAME + 1)
finally:
os.close(descriptor)
if len(raw) > MAX_FRAME:
raise ValueError("oversized promotion challenge")
challenge = raw.decode("utf-8")
if (
len(challenge) != 64
or any(character not in "0123456789abcdef" for character in challenge)
):
raise ValueError("invalid promotion challenge")
return PendingChallenge(challenge, metadata.st_dev, metadata.st_ino)
def write_result(
directory_descriptor: int,
attempt_id: str,
verified: bool,
reason: str | None,
session_id: str,
wall_clock: float,
) -> None:
result = {
"attempt_id": attempt_id,
"expires_at_wallclock": wall_clock + LEASE_TTL_SECONDS if verified else None,
"reason": reason,
"session_id": session_id,
"ts": wall_clock,
"verified": verified,
}
temporary = f".{RESULT_FILE}.tmp-{secrets.token_hex(8)}"
flags = (
os.O_WRONLY
| os.O_CREAT
| os.O_EXCL
| getattr(os, "O_CLOEXEC", 0)
| getattr(os, "O_NOFOLLOW", 0)
)
descriptor = os.open(temporary, flags, 0o600, dir_fd=directory_descriptor)
try:
os.fchmod(descriptor, 0o600)
with os.fdopen(descriptor, "w", encoding="utf-8", closefd=False) as stream:
json.dump(result, stream, separators=(",", ":"), sort_keys=True)
stream.flush()
os.fsync(stream.fileno())
os.replace(
temporary,
RESULT_FILE,
src_dir_fd=directory_descriptor,
dst_dir_fd=directory_descriptor,
)
os.fsync(directory_descriptor)
except Exception:
try:
os.unlink(temporary, dir_fd=directory_descriptor)
except FileNotFoundError:
pass
raise
finally:
os.close(descriptor)
def delete_pending_if_unchanged(
directory_descriptor: int,
name: str,
pending: PendingChallenge,
error_stream: TextIO,
) -> None:
quarantine = f".{name}.delete-{secrets.token_hex(8)}"
try:
os.rename(
name,
quarantine,
src_dir_fd=directory_descriptor,
dst_dir_fd=directory_descriptor,
)
except FileNotFoundError:
return
except OSError as error:
print(f"Mosaic promotion could not quarantine pending file: {error}", file=error_stream)
return
try:
moved = os.stat(
quarantine,
dir_fd=directory_descriptor,
follow_symlinks=False,
)
if (moved.st_dev, moved.st_ino) == (pending.device, pending.inode):
os.unlink(quarantine, dir_fd=directory_descriptor)
os.fsync(directory_descriptor)
return
print("Mosaic promotion pending file changed; preserving replacement.", file=error_stream)
try:
os.link(
quarantine,
name,
src_dir_fd=directory_descriptor,
dst_dir_fd=directory_descriptor,
follow_symlinks=False,
)
except FileExistsError:
print(
f"Mosaic promotion preserved replacement as {quarantine}.",
file=error_stream,
)
else:
os.unlink(quarantine, dir_fd=directory_descriptor)
os.fsync(directory_descriptor)
except OSError as error:
print(f"Mosaic promotion could not resolve pending file: {error}", file=error_stream)
def parse_reply(completed: subprocess.CompletedProcess[str]) -> dict[str, object] | None:
if completed.returncode != 0:
return None
try:
value = json.loads(
completed.stdout,
object_pairs_hook=reject_duplicate_json_keys,
)
except (json.JSONDecodeError, RecursionError, TypeError, ValueError):
return None
if not isinstance(value, dict):
return None
if set(value) == {"stage", "ok", "state"}:
if (
value.get("stage") == "promote_lease"
and value.get("ok") is True
and value.get("state") == "VERIFIED"
):
return value
return None
if set(value) == {"stage", "ok", "code"}:
if (
value.get("stage") in {"observe_receipt", "promote_lease"}
and value.get("ok") is False
and isinstance(value.get("code"), str)
and value.get("code")
):
return value
return None
def main(
*,
environ: Mapping[str, str] | None = None,
stderr: TextIO | None = None,
run: Callable[..., subprocess.CompletedProcess[str]] = subprocess.run,
now: Callable[[], float] = time.time,
) -> int:
source_environment = os.environ if environ is None else environ
error_stream = sys.stderr if stderr is None else stderr
directory_descriptor: int | None = None
lock_descriptor: int | None = None
try:
runtime_dir, pending_name = session_pending_name(source_environment)
session_id = source_environment["MOSAIC_LEASE_SESSION_ID"]
directory_descriptor = open_pending_directory(runtime_dir)
if directory_descriptor is None:
return 0
try:
lock_descriptor = acquire_lock(directory_descriptor)
except (BlockingIOError, FileNotFoundError):
print("Mosaic promotion completion deferred: promotion is in progress.", file=error_stream)
return 0
pending = read_pending(directory_descriptor, pending_name)
if pending is None:
return 0
completed = run(
[
sys.executable,
"-I",
"-S",
"-B",
str(PROMOTER),
"--complete",
pending.value,
],
check=False,
capture_output=True,
text=True,
env=dict(source_environment),
timeout=PROMOTER_TIMEOUT_SECONDS,
)
reply = parse_reply(completed)
if reply is not None and reply.get("ok") is True:
write_result(directory_descriptor, pending.value, True, None, session_id, now())
delete_pending_if_unchanged(
directory_descriptor,
pending_name,
pending,
error_stream,
)
print("Mosaic lease promotion completed.", file=error_stream)
return 0
if reply is not None:
code = str(reply["code"])
print(f"Mosaic promotion incomplete: {code}.", file=error_stream)
if code in TERMINAL_FAILURE_CODES:
write_result(directory_descriptor, pending.value, False, code, session_id, now())
delete_pending_if_unchanged(
directory_descriptor,
pending_name,
pending,
error_stream,
)
else:
diagnostic = completed.stderr.strip() or f"promoter exit {completed.returncode}"
print(f"Mosaic promotion retryable failure: {diagnostic}.", file=error_stream)
except (KeyError, OSError, RecursionError, UnicodeError, ValueError, subprocess.SubprocessError) as error:
print(f"Mosaic promotion completion deferred: {type(error).__name__}: {error}", file=error_stream)
finally:
if lock_descriptor is not None:
os.close(lock_descriptor)
if directory_descriptor is not None:
os.close(directory_descriptor)
return 0
if __name__ == "__main__":
raise SystemExit(main())
@@ -22,13 +22,23 @@ from typing import Final
MAX_FRAME: Final = 64 * 1024 MAX_FRAME: Final = 64 * 1024
BROKER_TIMEOUT_SECONDS: Final = 1.5 BROKER_TIMEOUT_SECONDS: Final = 1.5
MAX_TRANSCRIPT_BYTES: Final = 4 * 1024 * 1024 MAX_TRANSCRIPT_BYTES: Final = 4 * 1024 * 1024
BENIGN_OBSERVATION_UNAVAILABLE_CODE: Final = "OBSERVATION_UNAVAILABLE"
def reject_duplicate_json_keys(pairs: list[tuple[str, object]]) -> dict[str, object]:
value: dict[str, object] = {}
for key, item in pairs:
if key in value:
raise ValueError("duplicate observer JSON key")
value[key] = item
return value
def read_json(stream: object) -> dict[str, object]: def read_json(stream: object) -> dict[str, object]:
raw = getattr(stream, "buffer", stream).read(MAX_FRAME + 1) raw = getattr(stream, "buffer", stream).read(MAX_FRAME + 1)
if not isinstance(raw, bytes) or len(raw) > MAX_FRAME: if not isinstance(raw, bytes) or len(raw) > MAX_FRAME:
raise ValueError("invalid observer input") raise ValueError("invalid observer input")
value = json.loads(raw) value = json.loads(raw, object_pairs_hook=reject_duplicate_json_keys)
if not isinstance(value, dict): if not isinstance(value, dict):
raise ValueError("invalid observer input") raise ValueError("invalid observer input")
return value return value
@@ -100,9 +110,9 @@ def observer_request(socket_path: Path, request: dict[str, object]) -> dict[str,
if not chunk: if not chunk:
break break
response.extend(chunk) response.extend(chunk)
if len(response) > MAX_FRAME or not response.endswith(b"\n"): if len(response) > MAX_FRAME or response.count(b"\n") != 1 or not response.endswith(b"\n"):
raise ValueError("invalid observer reply") raise ValueError("invalid observer reply")
value = json.loads(response) value = json.loads(response[:-1], object_pairs_hook=reject_duplicate_json_keys)
if not isinstance(value, dict): if not isinstance(value, dict):
raise ValueError("invalid observer reply") raise ValueError("invalid observer reply")
return value return value
@@ -119,6 +129,11 @@ def main(argv: Sequence[str] | None = None, *, environ: Mapping[str, str] | None
if arguments.runtime == "claude": if arguments.runtime == "claude":
if not arguments.latest_entry: if not arguments.latest_entry:
raise ValueError("Claude observer requires --latest-entry") raise ValueError("Claude observer requires --latest-entry")
if "last_assistant_message" in source:
message = source["last_assistant_message"]
if not isinstance(message, str):
raise ValueError("invalid Claude observer input")
else:
message = claude_latest_entry(source) message = claude_latest_entry(source)
else: else:
if arguments.latest_entry: if arguments.latest_entry:
@@ -133,10 +148,18 @@ def main(argv: Sequence[str] | None = None, *, environ: Mapping[str, str] | None
"runtime": arguments.runtime, "runtime": arguments.runtime,
"latest_assistant_message": message, "latest_assistant_message": message,
}) })
except (KeyError, OSError, ValueError, json.JSONDecodeError) as error: except (KeyError, OSError, RecursionError, ValueError, json.JSONDecodeError) as error:
print(f"Mosaic receipt observer refused: {error}", file=sys.stderr) print(f"Mosaic receipt observer refused: {error}", file=sys.stderr)
return 2 return 2
return 0 if reply == {"ok": True} else 2 if set(reply) == {"ok"} and reply.get("ok") is True:
return 0
if (
set(reply) == {"ok", "code"}
and reply.get("ok") is False
and reply.get("code") == BENIGN_OBSERVATION_UNAVAILABLE_CODE
):
return 0
return 2
if __name__ == "__main__": if __name__ == "__main__":
@@ -33,7 +33,7 @@ def is_verbatim_receipt(message: str, challenge: str, binding: dict[str, object]
"""Require the exact one current-cycle receipt, not a transcript substring.""" """Require the exact one current-cycle receipt, not a transcript substring."""
expected = receipt_for(challenge, binding) expected = receipt_for(challenge, binding)
return hmac.compare_digest(message, expected) return hmac.compare_digest(message.encode("utf-8"), expected.encode("utf-8"))
def latest_assistant_digest(message: str) -> str: def latest_assistant_digest(message: str) -> str:
@@ -39,11 +39,12 @@ ORIG_PATH="$PATH"
# loop — which would make the control a false negative. A root dotfile is # loop — which would make the control a false negative. A root dotfile is
# operator-owned (unknown→operator), so the sync loop skips it. Clean up on exit. # operator-owned (unknown→operator), so the sync loop skips it. Clean up on exit.
STRIPPED="$FW/.install-rollback-control.tmp.sh" STRIPPED="$FW/.install-rollback-control.tmp.sh"
SIGNALED="$FW/.install-signal-control.tmp.sh"
NOEXIT="$FW/.install-noexit-control.tmp.sh" NOEXIT="$FW/.install-noexit-control.tmp.sh"
D1CTRL="$FW/.install-d1guard-control.tmp.sh" D1CTRL="$FW/.install-d1guard-control.tmp.sh"
D2CTRL="$FW/.install-d2guard-control.tmp.sh" D2CTRL="$FW/.install-d2guard-control.tmp.sh"
rm -f "$STRIPPED" "$NOEXIT" "$D1CTRL" "$D2CTRL" rm -f "$STRIPPED" "$SIGNALED" "$NOEXIT" "$D1CTRL" "$D2CTRL"
trap 'rm -f "$STRIPPED" "$NOEXIT" "$D1CTRL" "$D2CTRL"' EXIT trap 'rm -f "$STRIPPED" "$SIGNALED" "$NOEXIT" "$D1CTRL" "$D2CTRL"' EXIT
pass=0; fail=0 pass=0; fail=0
chk() { if eval "$2"; then echo "$1"; pass=$((pass + 1)); else echo "$1"; fail=$((fail + 1)); fi; } chk() { if eval "$2"; then echo "$1"; pass=$((pass + 1)); else echo "$1"; fail=$((fail + 1)); fi; }
@@ -180,41 +181,86 @@ chk "[control] without -E the mid-sync corruption survives (no rollback)" \
# ── Part C: an INT/TERM interrupt must terminate, not resume (blocker-A) ────── # ── Part C: an INT/TERM interrupt must terminate, not resume (blocker-A) ──────
# A bash signal trap that merely returns lets the script continue past the # A bash signal trap that merely returns lets the script continue past the
# interrupt — restoring the snapshot, then resuming the sync and reporting # interrupt — restoring the snapshot, then resuming the sync and reporting
# success. We inject a SIGTERM mid-sync with a cp that SUCCEEDS (so set -e never # success. The earlier test used a child cp shim to signal its parent, making
# fires and ONLY the signal path governs), and assert the shipped installer # child completion race Bash's interrupted wait. Concurrency is not part of the
# restores AND exits without reporting success. The control strips `exit 1` from # guarded property: sync_framework_keep() runs in the installer's own Bash
# the trap and shows the buggy resume-to-success. # process, and `kill` is a builtin. Generate two installer fixtures that signal
make_term_shim() { # themselves at the same known mid-sync point. Their TERM handlers emit the same
local dir="$1" # observable before diverging, so missing signal delivery fails BOTH arms rather
cat > "$dir/cp" <<SHIM # than manufacturing a pass. The only semantic difference between fixtures is
#!/usr/bin/env bash # the explicit `exit 1` whose load-bearing behavior this control proves.
dest="\${@: -1}" TERM_MARKER='[test-control] TERM handler entered'
case "\$dest" in HANDLER_WITH_EXIT="trap 'echo \"$TERM_MARKER\" >&2; restore_snapshot; exit 1' TERM # TEST-TERM-HANDLER"
*/$POISON_REL) HANDLER_WITHOUT_EXIT="trap 'echo \"$TERM_MARKER\" >&2; restore_snapshot' TERM # TEST-TERM-HANDLER"
kill -TERM "\$PPID" 2>/dev/null # signal install.sh; the copy still succeeds
exec env PATH="$ORIG_PATH" cp "\$@" ;; make_signal_installer() {
esac local output="$1" handler="$2"
exec env PATH="$ORIG_PATH" cp "\$@" local target_trap="trap 'restore_snapshot; exit 1' ERR INT TERM"
SHIM local target_cp=' cp "$abs" "$dst/$rel"'
chmod +x "$dir/cp" local inject_open=" if [[ \"\$rel\" == \"$POISON_REL\" ]]; then"
local inject_kill=' kill -TERM "$$" # TEST-TERM-INJECTION'
local inject_close=' fi'
if ! awk \
-v target_trap="$target_trap" -v target_cp="$target_cp" \
-v handler="$handler" -v inject_open="$inject_open" \
-v inject_kill="$inject_kill" -v inject_close="$inject_close" '
$0 == target_cp {
print inject_open
print inject_kill
print inject_close
injection_sites++
}
{ print }
$0 == target_trap {
print handler
handler_sites++
}
END {
if (handler_sites != 1 || injection_sites != 1) exit 42
}
' "$INSTALL" > "$output"; then
rm -f "$output"
fail "Could not construct the self-TERM control installer at the exact trap/copy sites"
exit 1
fi
chmod +x "$output"
} }
# Run one keep-mode upgrade with the SIGTERM shim. Echoes "<exit>\t<out>\t<home>". make_signal_installer "$SIGNALED" "$HANDLER_WITH_EXIT"
make_signal_installer "$NOEXIT" "$HANDLER_WITHOUT_EXIT"
signal_fixture_ready() {
local fixture="$1" expected_handler="$2"
[[ "$(grep -cF '# TEST-TERM-INJECTION' "$fixture")" -eq 1 ]] \
&& [[ "$(grep -cF '# TEST-TERM-HANDLER' "$fixture")" -eq 1 ]] \
&& grep -Fqx "$expected_handler" "$fixture"
}
signaled_fixture_ready() { signal_fixture_ready "$SIGNALED" "$HANDLER_WITH_EXIT"; }
noexit_fixture_ready() { signal_fixture_ready "$NOEXIT" "$HANDLER_WITHOUT_EXIT"; }
chk "[signal] shipped fixture has exactly one self-TERM injection and marked handler" \
"signaled_fixture_ready"
chk "[control] no-exit fixture has exactly one self-TERM injection and marked handler" \
"noexit_fixture_ready"
chk "[control] removing the explicit TERM exit changes the fixture" \
"! cmp -s '$SIGNALED' '$NOEXIT'"
# Run one keep-mode upgrade whose own shell delivers SIGTERM synchronously at
# the selected copy. Echoes "<exit>\t<out>\t<home>".
run_signal_upgrade() { run_signal_upgrade() {
local installer="$1" H OUT SHIM rc local installer="$1" H OUT rc
H=$(mktemp -d); OUT=$(mktemp); SHIM=$(mktemp -d) H=$(mktemp -d); OUT=$(mktemp)
seed_home "$H" seed_home "$H"
make_term_shim "$SHIM"
set +e set +e
PATH="$SHIM:$ORIG_PATH" \ PATH="$ORIG_PATH" \
MOSAIC_HOME="$H" MOSAIC_INSTALL_MODE=keep MOSAIC_SYNC_ONLY=1 bash "$installer" >"$OUT" 2>&1 MOSAIC_HOME="$H" MOSAIC_INSTALL_MODE=keep MOSAIC_SYNC_ONLY=1 bash "$installer" >"$OUT" 2>&1
rc=$? rc=$?
set -e 2>/dev/null || true set -e 2>/dev/null || true
rm -rf "$SHIM"
printf '%s\t%s\t%s\n' "$rc" "$OUT" "$H" printf '%s\t%s\t%s\n' "$rc" "$OUT" "$H"
} }
IFS=$'\t' read -r rcC OUTC HC < <(run_signal_upgrade "$INSTALL") IFS=$'\t' read -r rcC OUTC HC < <(run_signal_upgrade "$SIGNALED")
chk "[signal] TERM handler observable fires exactly once" \
"[ \"\$(grep -cF '$TERM_MARKER' '$OUTC')\" -eq 1 ]"
chk "[signal] SIGTERM mid-sync aborts non-zero (trap exits, does not resume)" \ chk "[signal] SIGTERM mid-sync aborts non-zero (trap exits, does not resume)" \
"[ '$rcC' -ne 0 ]" "[ '$rcC' -ne 0 ]"
chk "[signal] restore_snapshot fires on the interrupt" \ chk "[signal] restore_snapshot fires on the interrupt" \
@@ -222,13 +268,13 @@ chk "[signal] restore_snapshot fires on the interrupt" \
chk "[signal] does NOT resume to report sync success after the interrupt" \ chk "[signal] does NOT resume to report sync success after the interrupt" \
"! grep -q 'file phase complete' '$OUTC'" "! grep -q 'file phase complete' '$OUTC'"
# Control: strip `exit 1` from the signal trap → the handler returns, the script IFS=$'\t' read -r rcD OUTD HD < <(run_signal_upgrade "$NOEXIT")
# resumes past the interrupt and wrongly reports success. In $FW so SOURCE_DIR resolves. chk "[control] TERM handler observable fires exactly once" \
sed "s/trap 'restore_snapshot; exit 1' ERR INT TERM/trap 'restore_snapshot' ERR INT TERM/" \ "[ \"\$(grep -cF '$TERM_MARKER' '$OUTD')\" -eq 1 ]"
"$INSTALL" > "$NOEXIT" chk "[control] without 'exit 1' the handler restores before returning" \
chk "[control] the exit-strip actually changed the installer" \ "grep -q 'restoring previous state from snapshot' '$OUTD'"
"! cmp -s '$INSTALL' '$NOEXIT'" chk "[control] without 'exit 1' the installer exits zero after resuming" \
IFS=$'\t' read -r _rcD OUTD HD < <(run_signal_upgrade "$NOEXIT") "[ '$rcD' -eq 0 ]"
chk "[control] without 'exit 1' the trap resumes and reports sync success (the bug)" \ chk "[control] without 'exit 1' the trap resumes and reports sync success (the bug)" \
"grep -q 'file phase complete' '$OUTD'" "grep -q 'file phase complete' '$OUTD'"
@@ -309,10 +355,10 @@ chk "[control] without the D2 recovery line the operator gets no snapshot pointe
# Reap any snapshot the reset-fail runs left in /tmp (reset failed → never cleaned). # Reap any snapshot the reset-fail runs left in /tmp (reset failed → never cleaned).
grep -o '/[^ ]*mosaic-snapshot[^ ]*' "$OUTH" 2>/dev/null | head -1 | while read -r s; do rm -rf "$s"; done grep -o '/[^ ]*mosaic-snapshot[^ ]*' "$OUTH" 2>/dev/null | head -1 | while read -r s; do rm -rf "$s"; done
# Cleanup ($STRIPPED / $NOEXIT / $D1CTRL / $D2CTRL are also removed by the EXIT trap). # Cleanup (generated installer controls are also removed by the EXIT trap).
for d in "$HA" "$REFA" "$HB" "$REFB" "$HC" "$HD" "$HE" "$REFE" "$HF" "$REFF" "$HG" "$HH"; do rm -rf "$d"; done for d in "$HA" "$REFA" "$HB" "$REFB" "$HC" "$HD" "$HE" "$REFE" "$HF" "$REFF" "$HG" "$HH"; do rm -rf "$d"; done
rm -f "$OUTA" "$OUTB" "$OUTC" "$OUTD" "$OUTE" "$OUTF" "$OUTG" "$OUTH" \ rm -f "$OUTA" "$OUTB" "$OUTC" "$OUTD" "$OUTE" "$OUTF" "$OUTG" "$OUTH" \
"$STRIPPED" "$NOEXIT" "$D1CTRL" "$D2CTRL" "$STRIPPED" "$SIGNALED" "$NOEXIT" "$D1CTRL" "$D2CTRL"
echo echo
echo "RESULT: $pass passed, $fail failed" echo "RESULT: $pass passed, $fail failed"
+1 -1
View File
@@ -25,7 +25,7 @@
"lint": "eslint src", "lint": "eslint src",
"typecheck": "tsc --noEmit", "typecheck": "tsc --noEmit",
"test": "vitest run --passWithNoTests && pnpm run test:framework-shell", "test": "vitest run --passWithNoTests && pnpm run test:framework-shell",
"test:framework-shell": "bash framework/tools/quality/scripts/check-test-enumeration.sh && bash framework/tools/quality/scripts/test-check-test-enumeration.sh && python3 src/lease-broker/daemon_deadline_unittest.py && python3 src/lease-broker/normative_fragments_unittest.py && python3 src/lease-broker/receipt_challenge_unittest.py && python3 src/lease-broker/context_recovery_unittest.py && python3 src/lease-broker/recovery_runtime_unittest.py && python3 src/lease-broker/recovery_b1_adversarial_unittest.py && python3 src/lease-broker/framework_skill_portability_unittest.py && python3 src/mutator-gate/runtime_tools_unittest.py && python3 src/mutator-gate/runtime_launch_guard_unittest.py && python3 src/mutator-gate/version_coupling_unittest.py && python3 framework/tools/lease-broker/check-runtime-launches.py --root ../.. && bash framework/tools/codex/test-pr-diff-context.sh && bash framework/tools/qa/test-deps-preflight.sh && bash framework/tools/git/test-pr-review-gitea-comment.sh && bash framework/tools/git/test-pr-review-repo-host-override.sh && bash framework/tools/git/test-ci-queue-wait-branch-absent.sh && bash framework/tools/git/test-ci-queue-wait-tristate.sh && bash framework/tools/git/test-ci-queue-wait-github-checks.sh && bash framework/tools/git/test-pr-merge-queue-branch.sh && bash framework/tools/git/test-pr-merge-head-pin.sh && bash framework/tools/git/test-git-credential-mosaic.sh && bash framework/tools/git/test-gitea-token-identity.sh && bash framework/tools/woodpecker/test-terminal-green-contract.sh && bash framework/tools/_scripts/test-install-ordering-guard.sh && bash framework/tools/tmux/agent-send.test.sh && bash framework/tools/wake/test-wake-store-ack.sh && bash framework/tools/wake/test-wake-store-enqueue-race.sh && bash framework/tools/wake/test-wake-digest-hmac.sh && bash framework/tools/wake/test-wake-digest-quarantine.sh && bash framework/tools/wake/test-wake-detector.sh && bash framework/tools/wake/test-wake-fn-oracle.sh && bash framework/tools/wake/test-wake-reconcile.sh && bash framework/tools/wake/test-wake-beacon.sh && bash framework/tools/wake/test-wake-preimage.sh && bash framework/tools/wake/test-wake-install.sh" "test:framework-shell": "bash framework/tools/quality/scripts/check-test-enumeration.sh && bash framework/tools/quality/scripts/test-check-test-enumeration.sh && python3 src/lease-broker/daemon_deadline_unittest.py && python3 src/lease-broker/normative_fragments_unittest.py && python3 src/lease-broker/promotion_binding_unittest.py && python3 src/lease-broker/promotion_trigger_unittest.py && python3 src/lease-broker/receipt_challenge_unittest.py && python3 src/lease-broker/context_recovery_unittest.py && python3 src/lease-broker/recovery_runtime_unittest.py && python3 src/lease-broker/recovery_b1_adversarial_unittest.py && python3 src/lease-broker/receipt_observer_client_unittest.py && python3 src/lease-broker/invariant_r_unittest.py && python3 src/lease-broker/framework_skill_portability_unittest.py && python3 src/mutator-gate/runtime_tools_unittest.py && python3 src/mutator-gate/runtime_launch_guard_unittest.py && python3 src/mutator-gate/version_coupling_unittest.py && python3 framework/tools/lease-broker/check-runtime-launches.py --root ../.. && bash framework/tools/codex/test-pr-diff-context.sh && bash framework/tools/qa/test-deps-preflight.sh && bash framework/tools/git/test-pr-review-gitea-comment.sh && bash framework/tools/git/test-pr-review-repo-host-override.sh && bash framework/tools/git/test-ci-queue-wait-branch-absent.sh && bash framework/tools/git/test-ci-queue-wait-tristate.sh && bash framework/tools/git/test-ci-queue-wait-github-checks.sh && bash framework/tools/git/test-pr-merge-queue-branch.sh && bash framework/tools/git/test-pr-merge-head-pin.sh && bash framework/tools/git/test-pr-merge-message-field.sh && bash framework/tools/git/test-git-credential-mosaic.sh && bash framework/tools/git/test-gitea-token-identity.sh && bash framework/tools/woodpecker/test-terminal-green-contract.sh && bash framework/tools/_scripts/test-install-ordering-guard.sh && bash framework/tools/_scripts/test-mosaic-init-rce.sh && bash framework/tools/tmux/agent-send.test.sh && bash framework/tools/wake/test-wake-store-ack.sh && bash framework/tools/wake/test-wake-store-enqueue-race.sh && bash framework/tools/wake/test-wake-digest-hmac.sh && bash framework/tools/wake/test-wake-digest-quarantine.sh && bash framework/tools/wake/test-wake-detector.sh && bash framework/tools/wake/test-wake-fn-oracle.sh && bash framework/tools/wake/test-wake-reconcile.sh && bash framework/tools/wake/test-wake-beacon.sh && bash framework/tools/wake/test-wake-preimage.sh && bash framework/tools/wake/test-wake-install.sh"
}, },
"dependencies": { "dependencies": {
"@mosaicstack/brain": "workspace:*", "@mosaicstack/brain": "workspace:*",
+2
View File
@@ -15,6 +15,7 @@ import { registerAgentCommand } from './commands/agent.js';
import { registerInteractionCommand } from './commands/interaction.js'; import { registerInteractionCommand } from './commands/interaction.js';
import { registerConfigCommand } from './commands/config.js'; import { registerConfigCommand } from './commands/config.js';
import { registerFleetCommand } from './commands/fleet.js'; import { registerFleetCommand } from './commands/fleet.js';
import { registerPromoteCommand } from './commands/promote.js';
import { registerMissionCommand } from './commands/mission.js'; import { registerMissionCommand } from './commands/mission.js';
import { registerUninstallCommand } from './commands/uninstall.js'; import { registerUninstallCommand } from './commands/uninstall.js';
import { registerRestoreCommand } from './commands/restore.js'; import { registerRestoreCommand } from './commands/restore.js';
@@ -370,6 +371,7 @@ registerInteractionCommand(program);
// ─── fleet ───────────────────────────────────────────────────────────── // ─── fleet ─────────────────────────────────────────────────────────────
registerFleetCommand(program); registerFleetCommand(program);
registerPromoteCommand(program);
// ─── config ──────────────────────────────────────────────────────────── // ─── config ────────────────────────────────────────────────────────────
+175 -8
View File
@@ -14,9 +14,11 @@ import {
readdirSync, readdirSync,
realpathSync, realpathSync,
rmSync, rmSync,
appendFileSync,
} from 'node:fs'; } from 'node:fs';
import { createHash, randomBytes } from 'node:crypto';
import { createRequire } from 'node:module'; import { createRequire } from 'node:module';
import { homedir } from 'node:os'; import { homedir, hostname } from 'node:os';
import { join, dirname } from 'node:path'; import { join, dirname } from 'node:path';
import type { Command } from 'commander'; import type { Command } from 'commander';
import { import {
@@ -42,6 +44,163 @@ const RUNTIME_LABELS: Record<RuntimeName, string> = {
pi: 'Pi', pi: 'Pi',
}; };
// ─── Harness home isolation ──────────────────────────────────────────────────
// Mosaic-launched runtimes read config from a dedicated home under the mosaic
// tree — never the operator's base install. A bare `claude` / `pi` therefore
// keeps its own config AND its own auth, and stays a working break-glass no
// matter what mosaic does to its own tree.
//
// These paths are manifest-UNKNOWN, which resolves to operator ownership
// (framework-manifest.txt rule 3, #791), so a keep-mode `mosaic update` can
// neither overwrite nor prune them. Overwrite-mode install still would.
//
// opencode has no dedicated config-dir variable and follows XDG, so isolating it
// sets XDG_CONFIG_HOME for that process tree. That is blunter than the other
// three: it also relocates XDG lookups for anything opencode spawns.
const HARNESS_HOME_ENV: Record<RuntimeName, string> = {
claude: 'CLAUDE_CONFIG_DIR',
pi: 'PI_CODING_AGENT_DIR',
codex: 'CODEX_HOME',
opencode: 'XDG_CONFIG_HOME',
};
/** Dedicated mosaic-owned home for a runtime: ~/.config/mosaic/.<runtime> */
function harnessHome(runtime: RuntimeName): string {
return join(MOSAIC_HOME, `.${runtime}`);
}
/**
* Env overlay pointing a runtime at its mosaic-owned home. The directory is
* created on demand so a first launch does not fail on a missing path.
*/
function harnessEnv(runtime: RuntimeName): Record<string, string> {
const key = HARNESS_HOME_ENV[runtime];
if (!key) return {};
const home = harnessHome(runtime);
mkdirSync(home, { recursive: true });
return { [key]: home };
}
// ─── Launch record (immutable provenance) ────────────────────────────────────
// MANDATORY and MECHANICAL: every launch appends one record of what the agent
// actually launched with, written before exec. No model involvement, no opt-out.
//
// WHY LAUNCH-TIME AND NOT INSPECT-LATER: pi rewrites its own argv to a bare
// `pi`, so /proc/<pid>/cmdline DESTROYS the launch evidence. That has already
// produced a confident wrong diagnosis ("this agent bypassed the launcher"),
// disproved only by the parent process's argv and only because the parent had
// not yet exited. A record written before exec is the only place this survives.
//
// Lands in fleet/run/sessions/ — the #797 Runtime Session Ledger path, already
// operator-classified in framework-manifest.txt and already covered by
// test-upgrade-manifest-guard.sh, so an upgrade can neither overwrite nor prune
// it.
//
// CORRELATION is by an explicit MOSAIC_LAUNCH_ID, never by pid: execRuntime()
// uses spawnSync, so the runtime is a CHILD with a different pid.
// launch-runtime.py appends the matching `lease.register` event.
//
// NEVER records a credential value: env is captured as PRESENT NAMES ONLY, and
// oversized argv values (the composed system prompt) become a digest + length.
const LAUNCH_LEDGER_DIR = join(MOSAIC_HOME, 'fleet', 'run', 'sessions');
const CLI_VERSION: string | null = (() => {
try {
// Resolved RELATIVELY: the package `exports` map does not expose
// package.json, so '@mosaicstack/mosaic/package.json' throws
// ERR_PACKAGE_PATH_NOT_EXPORTED. Same relative depth from src/ and dist/.
return (createRequire(import.meta.url)('../../package.json') as { version: string }).version;
} catch {
return null;
}
})();
interface NormativeFragmentDigest {
source_id: string;
sha256: string | null;
bytes: number | null;
missing?: boolean;
}
function sha256Of(value: string | Buffer): string {
return createHash('sha256').update(value).digest('hex');
}
/**
* Hash the normative sources injected into the agent. This is "what the agent
* IS" and it is the same fragment set the lease broker hashes for promotion,
* so an unexpected digest here is a mechanically detectable red flag rather than
* a matter of judgement.
*/
function normativeFragmentDigests(runtime: RuntimeName): NormativeFragmentDigest[] {
const candidates: Array<[string, string]> = [
['CONSTITUTION.md', join(MOSAIC_HOME, 'CONSTITUTION.md')],
['AGENTS.md', join(MOSAIC_HOME, 'AGENTS.md')],
['SOUL.md', join(MOSAIC_HOME, 'SOUL.md')],
['USER.md', join(MOSAIC_HOME, 'USER.md')],
['STANDARDS.md', join(MOSAIC_HOME, 'STANDARDS.md')],
['TOOLS.md', join(MOSAIC_HOME, 'TOOLS.md')],
[`runtime/${runtime}/RUNTIME.md`, join(MOSAIC_HOME, 'runtime', runtime, 'RUNTIME.md')],
];
return candidates.map(([sourceId, path]) => {
try {
const bytes = readFileSync(path);
return { source_id: sourceId, sha256: sha256Of(bytes), bytes: bytes.length };
} catch {
return { source_id: sourceId, sha256: null, bytes: null, missing: true };
}
});
}
/** argv with oversized values replaced by a digest, so the record stays small
* and never inlines injected content verbatim. */
function redactArgv(argv: string[]): string[] {
return argv.map((a) =>
typeof a === 'string' && a.length > 256
? `<redacted sha256:${sha256Of(a).slice(0, 16)} bytes:${a.length}>`
: a,
);
}
function recordLaunch(runtime: RuntimeName, cliArgs: string[], yolo: boolean): void {
try {
mkdirSync(LAUNCH_LEDGER_DIR, { recursive: true, mode: 0o700 });
// Correlation id for the lease.register half. Set into process.env so it
// propagates through every `...process.env` / `...baseEnv` spread below.
const launchId = `${Date.now().toString(36)}-${randomBytes(6).toString('hex')}`;
process.env['MOSAIC_LAUNCH_ID'] = launchId;
const record = {
seq: Date.now(),
kind: 'session.launch',
launch_id: launchId,
ts: new Date().toISOString(),
host: hostname(),
pid: process.pid,
runtime,
mode: yolo ? 'yolo' : 'normal',
cwd: process.cwd(),
cli_version: CLI_VERSION,
config_home: harnessHome(runtime),
config_home_isolated: true,
config_home_env: HARNESS_HOME_ENV[runtime] ?? null,
argv: redactArgv(cliArgs),
normative_fragments: normativeFragmentDigests(runtime),
// names only — values are never recorded
mosaic_env_present: Object.keys(process.env)
.filter((k) => k.startsWith('MOSAIC_'))
.sort(),
};
appendFileSync(join(LAUNCH_LEDGER_DIR, 'events.ndjson'), `${JSON.stringify(record)}\n`, {
mode: 0o600,
});
} catch (err) {
// Never block a launch on bookkeeping — but never fail silently either.
console.error(
`[mosaic] WARNING: launch record not written: ${err instanceof Error ? err.message : String(err)}`,
);
}
}
// ─── Pre-flight checks ────────────────────────────────────────────────────── // ─── Pre-flight checks ──────────────────────────────────────────────────────
function checkMosaicHome(): void { function checkMosaicHome(): void {
@@ -105,11 +264,11 @@ interface SettingsAudit {
function auditClaudeSettings(): SettingsAudit { function auditClaudeSettings(): SettingsAudit {
const warnings: string[] = []; const warnings: string[] = [];
const settingsPath = join(homedir(), '.claude', 'settings.json'); const settingsPath = join(harnessHome('claude'), 'settings.json');
const settings = readJson(settingsPath); const settings = readJson(settingsPath);
if (!settings) { if (!settings) {
warnings.push('~/.claude/settings.json not found — hooks and plugins will be missing'); warnings.push(`${settingsPath} not found — hooks and plugins will be missing`);
return { warnings }; return { warnings };
} }
@@ -561,7 +720,9 @@ function skillRealPath(dir: string): string {
/** Skill roots Pi auto-discovers natively (no `--skill` needed): its global /** Skill roots Pi auto-discovers natively (no `--skill` needed): its global
* skills dir and the project-local one relative to the launch cwd. */ * skills dir and the project-local one relative to the launch cwd. */
function piNativeSkillRoots(cwd: string = process.cwd()): string[] { function piNativeSkillRoots(cwd: string = process.cwd()): string[] {
return [join(homedir(), '.pi', 'agent', 'skills'), join(cwd, '.pi', 'skills')]; // PI_CODING_AGENT_DIR replaces ~/.pi/agent (not ~/.pi), so skills live at
// <home>/skills — there is no extra 'agent' segment under the isolated home.
return [join(harnessHome('pi'), 'skills'), join(cwd, '.pi', 'skills')];
} }
/** Enumerate skill dirs under a set of roots, deduped by real path. A directory /** Enumerate skill dirs under a set of roots, deduped by real path. A directory
@@ -764,12 +925,13 @@ function launchRuntime(runtime: RuntimeName, args: string[], yolo: boolean): nev
cliArgs.push(...args); cliArgs.push(...args);
} }
console.log(`[mosaic] Launching ${label}${modeStr}${missionStr}...`); console.log(`[mosaic] Launching ${label}${modeStr}${missionStr}...`);
recordLaunch('claude', cliArgs, yolo);
execLeaseGatedRuntime('claude', cliArgs, process.env, yolo); execLeaseGatedRuntime('claude', cliArgs, process.env, yolo);
break; break;
} }
case 'codex': { case 'codex': {
ensureRuntimeConfig('codex', join(homedir(), '.codex', 'instructions.md')); ensureRuntimeConfig('codex', join(harnessHome('codex'), 'instructions.md'));
const cliArgs = yolo ? ['--dangerously-bypass-approvals-and-sandbox'] : []; const cliArgs = yolo ? ['--dangerously-bypass-approvals-and-sandbox'] : [];
if (hasMissionNoArgs) { if (hasMissionNoArgs) {
cliArgs.push(missionPrompt); cliArgs.push(missionPrompt);
@@ -777,14 +939,17 @@ function launchRuntime(runtime: RuntimeName, args: string[], yolo: boolean): nev
cliArgs.push(...args); cliArgs.push(...args);
} }
console.log(`[mosaic] Launching ${label}${modeStr}${missionStr}...`); console.log(`[mosaic] Launching ${label}${modeStr}${missionStr}...`);
execRuntime('codex', cliArgs); recordLaunch('codex', cliArgs, yolo);
execRuntime('codex', cliArgs, { ...process.env, ...harnessEnv('codex') });
break; break;
} }
case 'opencode': { case 'opencode': {
ensureRuntimeConfig('opencode', join(homedir(), '.config', 'opencode', 'AGENTS.md')); // opencode follows XDG, so its config resolves to $XDG_CONFIG_HOME/opencode.
ensureRuntimeConfig('opencode', join(harnessHome('opencode'), 'opencode', 'AGENTS.md'));
console.log(`[mosaic] Launching ${label}${modeStr}...`); console.log(`[mosaic] Launching ${label}${modeStr}...`);
execRuntime('opencode', args); recordLaunch('opencode', args, yolo);
execRuntime('opencode', args, { ...process.env, ...harnessEnv('opencode') });
break; break;
} }
@@ -799,6 +964,7 @@ function launchRuntime(runtime: RuntimeName, args: string[], yolo: boolean): nev
cliArgs.push(...args); cliArgs.push(...args);
} }
console.log(`[mosaic] Launching ${label}${modeStr}${missionStr}...`); console.log(`[mosaic] Launching ${label}${modeStr}${missionStr}...`);
recordLaunch('pi', cliArgs, yolo);
execLeaseGatedRuntime('pi', cliArgs); execLeaseGatedRuntime('pi', cliArgs);
break; break;
} }
@@ -835,6 +1001,7 @@ function execLeaseGatedRuntime(
[launcher, ...dangerousArgs, '--runtime', runtime, '--', runtime, ...args], [launcher, ...dangerousArgs, '--runtime', runtime, '--', runtime, ...args],
{ {
...baseEnv, ...baseEnv,
...harnessEnv(runtime),
MOSAIC_LEASE_BROKER_SOCKET: defaultLeaseBrokerSocket(baseEnv), MOSAIC_LEASE_BROKER_SOCKET: defaultLeaseBrokerSocket(baseEnv),
MOSAIC_RUNTIME_GENERATION: baseEnv['MOSAIC_RUNTIME_GENERATION'] ?? '1', MOSAIC_RUNTIME_GENERATION: baseEnv['MOSAIC_RUNTIME_GENERATION'] ?? '1',
}, },
@@ -0,0 +1,298 @@
import { Command } from 'commander';
import { describe, expect, it, vi } from 'vitest';
import type { FleetRoster } from './fleet.js';
import { TmuxPromotionTransport } from '../fleet/promotion-transport.js';
import {
promoteSeat,
registerPromoteCommand,
type PromotionBreadcrumbStore,
type PromotionResult,
type PromotionTransport,
} from './promote.js';
const attemptId = 'a'.repeat(64);
const target = {
bundle: 'local',
seat: 'claude-seat',
sessionId: 'b'.repeat(64),
};
function transport(): PromotionTransport {
return {
resolve: vi.fn(async () => target),
sendPromotion: vi.fn(async () => {}),
};
}
describe('mosaic promote', () => {
it('accepts only a fresh result correlated to this attempt', async () => {
const promotionTransport = transport();
const store: PromotionBreadcrumbStore = {
readAttemptId: vi.fn().mockResolvedValueOnce(null).mockResolvedValue(attemptId),
readResult: vi
.fn()
.mockResolvedValueOnce({
attempt_id: 'c'.repeat(64),
expires_at_wallclock: 4_600,
reason: null,
session_id: target.sessionId,
ts: 1_001,
verified: true,
})
.mockResolvedValueOnce({
attempt_id: attemptId,
expires_at_wallclock: 4_600,
reason: null,
session_id: target.sessionId,
ts: 1_001,
verified: true,
}),
};
const result = await promoteSeat('claude-seat', {
clock: () => 1_000,
sleep: async () => {},
store,
timeoutMs: 1,
transport: promotionTransport,
});
expect(result).toEqual({
bundle: 'local',
expiresAtWallclock: 4_600,
reason: null,
seat: 'claude-seat',
sessionId: 'b'.repeat(64),
status: 'VERIFIED',
});
expect(promotionTransport.sendPromotion).toHaveBeenCalledWith(target);
expect(store.readResult).toHaveBeenCalledTimes(2);
});
it('accepts a fresh result for this session when completion consumed the pending nonce', async () => {
const promotionTransport = transport();
let now = 1_000;
const store: PromotionBreadcrumbStore = {
readAttemptId: vi.fn(async () => null),
readResult: vi.fn(async () => ({
attempt_id: attemptId,
expires_at_wallclock: 4_600,
reason: null,
session_id: target.sessionId,
ts: 1_001,
verified: true,
})),
};
const result = await promoteSeat('claude-seat', {
clock: () => now,
sleep: async () => {
now += 10;
},
store,
timeoutMs: 10,
transport: promotionTransport,
});
expect(result.status).toBe('VERIFIED');
});
it('returns UNVERIFIED within the command bound when a tmux runner wedges', async () => {
vi.useFakeTimers();
const roster: FleetRoster = {
agents: [{ className: 'worker', name: 'claude-seat', runtime: 'claude' }],
defaults: { workingDirectory: '~/src' },
runtimes: {},
tmux: { holderSession: '_holder', socketName: 'mosaic-fleet' },
transport: 'tmux',
version: 1,
};
const promotionTransport = new TmuxPromotionTransport({
mosaicHome: '/mosaic',
rosterLoader: async () => roster,
runner: async () => new Promise(() => {}),
});
const store: PromotionBreadcrumbStore = {
readAttemptId: vi.fn(async () => null),
readResult: vi.fn(async () => null),
};
try {
let observedResult: PromotionResult | undefined;
void promoteSeat('claude-seat', {
store,
transport: promotionTransport,
}).then((result) => {
observedResult = result;
});
await vi.advanceTimersByTimeAsync(5_000);
expect(observedResult).toMatchObject({
reason: 'RESOLVE_FAILED: Promotion transport command timed out after 5000ms.',
seat: 'claude-seat',
status: 'UNVERIFIED',
});
} finally {
vi.useRealTimers();
}
});
it('prints VERIFIED with the resolved seat, session, bundle, and wall-clock expiry', async () => {
const promotionTransport = transport();
const store: PromotionBreadcrumbStore = {
readAttemptId: vi.fn().mockResolvedValueOnce(null).mockResolvedValue(attemptId),
readResult: vi.fn(async () => ({
attempt_id: attemptId,
expires_at_wallclock: 4_600,
reason: null,
session_id: target.sessionId,
ts: Number.MAX_SAFE_INTEGER,
verified: true,
})),
};
const output = vi.spyOn(console, 'log').mockImplementation(() => {});
const program = new Command().exitOverride();
registerPromoteCommand(program, {
mintAuthorization: async () => {},
store,
transport: promotionTransport,
});
try {
await program.parseAsync(['node', 'mosaic', 'promote', 'claude-seat']);
expect(output).toHaveBeenCalledWith(
`VERIFIED seat=claude-seat session=${target.sessionId} bundle=local expiry=1970-01-01T01:16:40.000Z`,
);
expect(process.exitCode).not.toBe(1);
} finally {
output.mockRestore();
process.exitCode = undefined;
}
});
it('prints UNVERIFIED and exits 1 when delivery fails', async () => {
const promotionTransport: PromotionTransport = {
resolve: vi.fn(async () => target),
sendPromotion: vi.fn(async () => {
throw new Error('tmux unavailable');
}),
};
const store: PromotionBreadcrumbStore = {
readAttemptId: vi.fn(async () => null),
readResult: vi.fn(async () => null),
};
const output = vi.spyOn(console, 'log').mockImplementation(() => {});
const program = new Command().exitOverride();
registerPromoteCommand(program, {
mintAuthorization: async () => {},
store,
transport: promotionTransport,
});
try {
process.exitCode = undefined;
await program.parseAsync(['node', 'mosaic', 'promote', 'claude-seat']);
expect(output).toHaveBeenCalledWith(
`UNVERIFIED seat=claude-seat session=${target.sessionId} bundle=local expiry=none reason=DELIVERY_FAILED: tmux unavailable`,
);
expect(process.exitCode).toBe(1);
} finally {
output.mockRestore();
process.exitCode = undefined;
}
});
it('rejects a stale result even when its nonce matches', async () => {
const promotionTransport = transport();
let now = 1_000;
const store: PromotionBreadcrumbStore = {
readAttemptId: vi.fn().mockResolvedValueOnce(null).mockResolvedValue(attemptId),
readResult: vi.fn(async () => ({
attempt_id: attemptId,
expires_at_wallclock: 4_600,
reason: null,
session_id: target.sessionId,
ts: 1_000,
verified: true,
})),
};
const result = await promoteSeat('claude-seat', {
clock: () => now,
sleep: async () => {
now += 10;
},
store,
timeoutMs: 10,
transport: promotionTransport,
});
expect(result).toEqual({
bundle: 'local',
expiresAtWallclock: null,
reason: 'PROMOTION_TIMEOUT',
seat: 'claude-seat',
sessionId: 'b'.repeat(64),
status: 'UNVERIFIED',
});
});
it('does not accept a result for a pending attempt that existed before send', async () => {
const promotionTransport = transport();
let now = 1_000;
const store: PromotionBreadcrumbStore = {
readAttemptId: vi.fn(async () => attemptId),
readResult: vi.fn(async () => ({
attempt_id: attemptId,
expires_at_wallclock: 4_600,
reason: null,
session_id: target.sessionId,
ts: 1_001,
verified: true,
})),
};
const result = await promoteSeat('claude-seat', {
clock: () => now,
sleep: async () => {
now += 10;
},
store,
timeoutMs: 10,
transport: promotionTransport,
});
expect(result.status).toBe('UNVERIFIED');
expect(store.readResult).not.toHaveBeenCalled();
});
it('returns UNVERIFIED after a bounded timeout without reading stdin', async () => {
const promotionTransport = transport();
let now = 1_000;
const store: PromotionBreadcrumbStore = {
readAttemptId: vi.fn(async () => null),
readResult: vi.fn(async () => null),
};
const result = await promoteSeat('claude-seat', {
clock: () => now,
sleep: async () => {
now += 10;
},
store,
timeoutMs: 10,
transport: promotionTransport,
});
expect(result).toEqual({
bundle: 'local',
expiresAtWallclock: null,
reason: 'PROMOTION_TIMEOUT',
seat: 'claude-seat',
sessionId: 'b'.repeat(64),
status: 'UNVERIFIED',
});
expect(promotionTransport.sendPromotion).toHaveBeenCalledOnce();
expect(store.readResult).toHaveBeenCalledTimes(2);
});
});
+369
View File
@@ -0,0 +1,369 @@
import { spawn } from 'node:child_process';
import { randomBytes } from 'node:crypto';
import { constants } from 'node:fs';
import { mkdir, open, rename } from 'node:fs/promises';
import { join } from 'node:path';
import type { Command } from 'commander';
import {
TmuxPromotionTransport,
type PromotionTarget,
type PromotionTransport,
} from '../fleet/promotion-transport.js';
export type { PromotionTransport } from '../fleet/promotion-transport.js';
import { resolveFleetPaths, type CommandRunner } from './fleet.js';
const ATTEMPT_ID_PATTERN = /^[a-f0-9]{64}$/;
const DEFAULT_POLL_INTERVAL_MS = 250;
const DEFAULT_TIMEOUT_MS = 30_000;
const SUBPROCESS_TIMEOUT_MS = 4_500;
const PENDING_DIRECTORY = 'mosaic-lease';
const AUTHORIZATION_DIRECTORY = 'authorizations';
const AUTHORIZATION_TTL_SECONDS = 60;
const RESULT_FILE = 'last-result.json';
export interface PromotionBreadcrumb {
attempt_id: string;
expires_at_wallclock: number | null;
reason: string | null;
session_id: string;
ts: number;
verified: boolean;
}
export interface PromotionBreadcrumbStore {
readAttemptId(sessionId: string): Promise<string | null>;
readResult(): Promise<PromotionBreadcrumb | null>;
}
export interface PromotionResult {
bundle: string;
expiresAtWallclock: number | null;
reason: string | null;
seat: string;
sessionId: string;
status: 'VERIFIED' | 'UNVERIFIED';
}
export interface PromoteSeatOptions {
clock?: () => number;
pollIntervalMs?: number;
sleep?: (milliseconds: number) => Promise<void>;
store: PromotionBreadcrumbStore;
target?: PromotionTarget;
timeoutMs?: number;
transport: PromotionTransport;
}
export interface PromoteCommandDeps {
mintAuthorization?: (target: PromotionTarget) => Promise<void>;
mosaicHome?: string;
runner?: CommandRunner;
store?: PromotionBreadcrumbStore;
transport?: PromotionTransport;
}
/** Private, local result store shared with the in-seat completion hook. */
export class FilePromotionBreadcrumbStore implements PromotionBreadcrumbStore {
constructor(private readonly runtimeDirectory = defaultRuntimeDirectory()) {}
async readAttemptId(sessionId: string): Promise<string | null> {
if (!ATTEMPT_ID_PATTERN.test(sessionId)) return null;
const content = await readPrivateFile(
join(this.runtimeDirectory, PENDING_DIRECTORY, `pending-${sessionId}`),
);
const attemptId = content?.trim();
return attemptId !== undefined && ATTEMPT_ID_PATTERN.test(attemptId) ? attemptId : null;
}
async readResult(): Promise<PromotionBreadcrumb | null> {
const content = await readPrivateFile(
join(this.runtimeDirectory, PENDING_DIRECTORY, RESULT_FILE),
);
if (content === null) return null;
try {
return parseBreadcrumb(JSON.parse(content) as unknown);
} catch {
return null;
}
}
}
/** Drives one bounded, non-interactive in-seat promotion attempt. */
export async function promoteSeat(
seat: string,
options: PromoteSeatOptions,
): Promise<PromotionResult> {
const clock = options.clock ?? wallClockSeconds;
const sleep = options.sleep ?? defaultSleep;
const timeoutMs = normalizeTimeout(options.timeoutMs);
const pollIntervalMs = normalizePollInterval(options.pollIntervalMs);
let target: PromotionTarget;
try {
target = options.target ?? (await options.transport.resolve(seat));
} catch (error: unknown) {
return unverifiedUnresolvedSeat(seat, `RESOLVE_FAILED: ${errorMessage(error)}`);
}
const previousAttemptId = await options.store.readAttemptId(target.sessionId);
const preSendTimestamp = clock();
try {
await options.transport.sendPromotion(target);
} catch (error: unknown) {
return unverified(target, `DELIVERY_FAILED: ${errorMessage(error)}`);
}
const deadline = preSendTimestamp + timeoutMs / 1_000;
let attemptId: string | null = null;
while (true) {
const currentAttemptId = await options.store.readAttemptId(target.sessionId);
if (currentAttemptId !== null && currentAttemptId !== previousAttemptId) {
attemptId = currentAttemptId;
}
if (attemptId !== null || previousAttemptId === null) {
// Completion can consume a first attempt's nonce before this poll observes it.
// In that branch, correlation degrades to session_id + fresh timestamp, which
// is acceptable for this 0600, same-UID local trust boundary.
const breadcrumb = await options.store.readResult();
if (
breadcrumb !== null &&
breadcrumb.session_id === target.sessionId &&
(attemptId === null || breadcrumb.attempt_id === attemptId) &&
breadcrumb.ts > preSendTimestamp
) {
return {
bundle: target.bundle,
expiresAtWallclock: breadcrumb.expires_at_wallclock,
reason: breadcrumb.reason,
seat: target.seat,
sessionId: target.sessionId,
status: breadcrumb.verified ? 'VERIFIED' : 'UNVERIFIED',
};
}
}
if (clock() >= deadline) return unverified(target, 'PROMOTION_TIMEOUT');
await sleep(Math.min(pollIntervalMs, Math.max(0, deadline - clock()) * 1_000));
}
}
export function registerPromoteCommand(program: Command, deps: PromoteCommandDeps = {}): void {
const mosaicHome = deps.mosaicHome ?? resolveFleetPaths().mosaicHome;
const transport =
deps.transport ?? new TmuxPromotionTransport({ mosaicHome, runner: deps.runner ?? runCommand });
const store = deps.store ?? new FilePromotionBreadcrumbStore();
program
.command('promote <seat>')
.description('Promote a Claude fleet seat and report the correlated lease result')
.option(
'--timeout <ms>',
`Bounded result wait in milliseconds (default: ${DEFAULT_TIMEOUT_MS})`,
)
.action(async (seat: string, opts: { timeout?: string }) => {
if (
process.env['MOSAIC_LEASE_SESSION_ID'] !== undefined &&
deps.mintAuthorization === undefined
) {
console.error('mosaic promote must run outside a lease-gated seat.');
process.exitCode = 1;
return;
}
let target: PromotionTarget;
try {
target = await transport.resolve(seat);
await (deps.mintAuthorization ?? mintAuthorization)(target);
} catch (error: unknown) {
console.error(`mosaic promote authorization failed: ${errorMessage(error)}`);
process.exitCode = 1;
return;
}
const result = await promoteSeat(seat, {
store,
target,
timeoutMs: parseOptionTimeout(opts.timeout),
transport,
});
const expiry =
result.expiresAtWallclock === null
? 'none'
: new Date(result.expiresAtWallclock * 1_000).toISOString();
const reason = result.reason === null ? '' : ` reason=${result.reason}`;
console.log(
`${result.status} seat=${result.seat} session=${result.sessionId} bundle=${result.bundle} expiry=${expiry}${reason}`,
);
if (result.status === 'UNVERIFIED') process.exitCode = 1;
});
}
async function mintAuthorization(target: PromotionTarget): Promise<void> {
const directory = join(defaultRuntimeDirectory(), PENDING_DIRECTORY, AUTHORIZATION_DIRECTORY);
await mkdir(directory, { mode: 0o700, recursive: true });
const token = {
expires_at: wallClockSeconds() + AUTHORIZATION_TTL_SECONDS,
nonce: randomBytes(32).toString('hex'),
seat: target.seat,
session_id: target.sessionId,
ts: wallClockSeconds(),
};
const destination = join(directory, `${target.sessionId}.auth`);
const temporary = join(directory, `.${target.sessionId}.${randomBytes(8).toString('hex')}.tmp`);
const handle = await open(
temporary,
constants.O_WRONLY | constants.O_CREAT | constants.O_EXCL,
0o600,
);
try {
await handle.chmod(0o600);
await handle.writeFile(JSON.stringify(token));
await handle.sync();
} finally {
await handle.close();
}
await rename(temporary, destination);
}
function defaultRuntimeDirectory(): string {
const configured = process.env['XDG_RUNTIME_DIR'];
if (configured) return configured;
const uid = typeof process.getuid === 'function' ? process.getuid() : 0;
return `/run/user/${uid}`;
}
async function readPrivateFile(path: string): Promise<string | null> {
let handle: Awaited<ReturnType<typeof open>>;
try {
handle = await open(path, constants.O_RDONLY | constants.O_NOFOLLOW);
} catch {
return null;
}
try {
const metadata = await handle.stat();
if (
!metadata.isFile() ||
metadata.uid !== (typeof process.getuid === 'function' ? process.getuid() : 0) ||
(metadata.mode & 0o077) !== 0
) {
return null;
}
return handle.readFile({ encoding: 'utf8' });
} catch {
return null;
} finally {
await handle.close();
}
}
function parseBreadcrumb(value: unknown): PromotionBreadcrumb | null {
if (!isRecord(value) || Object.keys(value).length !== 6) return null;
const { attempt_id, expires_at_wallclock, reason, session_id, ts, verified } = value;
if (
typeof attempt_id !== 'string' ||
!ATTEMPT_ID_PATTERN.test(attempt_id) ||
typeof verified !== 'boolean' ||
typeof session_id !== 'string' ||
!ATTEMPT_ID_PATTERN.test(session_id) ||
typeof ts !== 'number' ||
!Number.isFinite(ts) ||
(expires_at_wallclock !== null &&
(typeof expires_at_wallclock !== 'number' || !Number.isFinite(expires_at_wallclock))) ||
(reason !== null && typeof reason !== 'string')
) {
return null;
}
return { attempt_id, expires_at_wallclock, reason, session_id, ts, verified };
}
function isRecord(value: unknown): value is Record<string, unknown> {
return typeof value === 'object' && value !== null && !Array.isArray(value);
}
function normalizeTimeout(value: number | undefined): number {
return value !== undefined && Number.isFinite(value) ? Math.max(0, value) : DEFAULT_TIMEOUT_MS;
}
function normalizePollInterval(value: number | undefined): number {
return value !== undefined && Number.isFinite(value)
? Math.max(1, value)
: DEFAULT_POLL_INTERVAL_MS;
}
function parseOptionTimeout(value: string | undefined): number | undefined {
if (value === undefined) return undefined;
const parsed = Number.parseInt(value, 10);
return Number.isFinite(parsed) ? parsed : undefined;
}
function unverifiedUnresolvedSeat(seat: string, reason: string): PromotionResult {
return {
bundle: 'unresolved',
expiresAtWallclock: null,
reason,
seat,
sessionId: 'unresolved',
status: 'UNVERIFIED',
};
}
function unverified(target: PromotionTarget, reason: string): PromotionResult {
return {
bundle: target.bundle,
expiresAtWallclock: null,
reason,
seat: target.seat,
sessionId: target.sessionId,
status: 'UNVERIFIED',
};
}
function errorMessage(error: unknown): string {
return error instanceof Error ? error.message : String(error);
}
function wallClockSeconds(): number {
return Date.now() / 1_000;
}
function defaultSleep(milliseconds: number): Promise<void> {
return new Promise((resolve) => setTimeout(resolve, milliseconds));
}
function runCommand(
command: string,
args: string[],
): Promise<{
exitCode: number;
stderr: string;
stdout: string;
}> {
return new Promise((resolve) => {
const child = spawn(command, args, { stdio: ['ignore', 'pipe', 'pipe'] });
let stdout = '';
let stderr = '';
let settled = false;
const finish = (result: { exitCode: number; stderr: string; stdout: string }): void => {
if (settled) return;
settled = true;
clearTimeout(timeout);
resolve(result);
};
const timeout = setTimeout(() => {
child.kill('SIGKILL');
finish({
exitCode: 124,
stderr: `Promotion transport subprocess timed out after ${SUBPROCESS_TIMEOUT_MS}ms.`,
stdout,
});
}, SUBPROCESS_TIMEOUT_MS);
child.stdout.on('data', (chunk: Buffer) => {
stdout += chunk.toString('utf8');
});
child.stderr.on('data', (chunk: Buffer) => {
stderr += chunk.toString('utf8');
});
child.on('error', (error: Error) => {
finish({ exitCode: 127, stderr: error.message, stdout });
});
child.on('close', (code: number | null) => {
finish({ exitCode: code ?? 1, stderr, stdout });
});
});
}
+8 -3
View File
@@ -332,9 +332,11 @@ describe('Claude skill bridge', () => {
expect(result.status, result.stderr).toBe(0); expect(result.status, result.stderr).toBe(0);
expect(readlinkSync(liveForeignLink)).toBe(liveForeignTarget); expect(readlinkSync(liveForeignLink)).toBe(liveForeignTarget);
expect(readlinkSync(danglingForeignLink)).toBe(join(mosaicHome, 'foreign-missing')); expect(readlinkSync(danglingForeignLink)).toBe(join(mosaicHome, 'foreign-missing'));
expect(readlinkSync(join(paths.claudeSkillsDir, 'missing'))).toBe( // Skills now link into the mosaic-owned harness home, never the base install.
expect(readlinkSync(join(mosaicHome, '.claude', 'skills', 'missing'))).toBe(
join(paths.mosaicSkillsDir, 'missing'), join(paths.mosaicSkillsDir, 'missing'),
); );
expect(existsSync(join(paths.claudeSkillsDir, 'missing'))).toBe(false);
}); });
it('preserves live and dangling foreign Claude symlinks while linking missing skills', () => { it('preserves live and dangling foreign Claude symlinks while linking missing skills', () => {
@@ -349,17 +351,20 @@ describe('Claude skill bridge', () => {
symlinkSync(external, liveLink); symlinkSync(external, liveLink);
symlinkSync(join(root, 'external-missing'), danglingLink); symlinkSync(join(root, 'external-missing'), danglingLink);
const mosaicHome = join(root, '.config', 'mosaic');
const result = spawnSync('bash', [LEGACY_SYNC_SCRIPT, '--link-only'], { const result = spawnSync('bash', [LEGACY_SYNC_SCRIPT, '--link-only'], {
encoding: 'utf8', encoding: 'utf8',
env: { ...process.env, HOME: root, MOSAIC_HOME: join(root, '.config', 'mosaic') }, env: { ...process.env, HOME: root, MOSAIC_HOME: mosaicHome },
}); });
expect(result.status, result.stderr).toBe(0); expect(result.status, result.stderr).toBe(0);
expect(readlinkSync(liveLink)).toBe(external); expect(readlinkSync(liveLink)).toBe(external);
expect(readlinkSync(danglingLink)).toBe(join(root, 'external-missing')); expect(readlinkSync(danglingLink)).toBe(join(root, 'external-missing'));
expect(readlinkSync(join(paths.claudeSkillsDir, 'missing'))).toBe( // Skills now link into the mosaic-owned harness home, never the base install.
expect(readlinkSync(join(mosaicHome, '.claude', 'skills', 'missing'))).toBe(
join(paths.mosaicSkillsDir, 'missing'), join(paths.mosaicSkillsDir, 'missing'),
); );
expect(existsSync(join(paths.claudeSkillsDir, 'missing'))).toBe(false);
}); });
}); });
@@ -0,0 +1,104 @@
import { describe, expect, it, vi } from 'vitest';
import type { CommandResult, CommandRunner, FleetRoster } from '../commands/fleet.js';
import { TmuxPromotionTransport } from './promotion-transport.js';
const sessionId = 'a'.repeat(64);
const roster: FleetRoster = {
agents: [{ className: 'worker', name: 'claude-seat', runtime: 'claude' }],
defaults: { workingDirectory: '~/src' },
runtimes: {},
tmux: { holderSession: '_holder', socketName: 'mosaic-fleet' },
transport: 'tmux',
version: 1,
};
function result(stdout = '', exitCode = 0, stderr = ''): CommandResult {
return { exitCode, stderr, stdout };
}
describe('TmuxPromotionTransport', () => {
it('resolves the exact roster seat and sends the registered command literally', async () => {
const runner = vi
.fn<CommandRunner>()
.mockResolvedValueOnce(result('1234 claude 0 0 0 0\n'))
.mockResolvedValueOnce(result())
.mockResolvedValueOnce(result());
const environmentReader = vi.fn(async () => `MOSAIC_LEASE_SESSION_ID=${sessionId}\0`);
const transport = new TmuxPromotionTransport({
environmentReader,
mosaicHome: '/mosaic',
rosterLoader: async () => roster,
runner,
});
const target = await transport.resolve('claude-seat');
await transport.sendPromotion(target);
expect(target).toEqual({
bundle: 'mosaic-fleet',
seat: 'claude-seat',
sessionId,
});
expect(environmentReader).toHaveBeenCalledWith(1234);
expect(runner).toHaveBeenNthCalledWith(2, 'tmux', [
'-L',
'mosaic-fleet',
'send-keys',
'-t',
'=claude-seat:0.0',
'-l',
'/mosaic-promote',
]);
expect(runner).toHaveBeenNthCalledWith(3, 'tmux', [
'-L',
'mosaic-fleet',
'send-keys',
'-t',
'=claude-seat:0.0',
'Enter',
]);
});
// Regression for #1124: the launcher runs claude as a spawnSync CHILD of
// node(mosaic), so the lease env is on the child, not the tmux pane pid. The
// transport must WALK the subtree. This test exercises the real walk (no
// full mock of the resolution) — the seam the original unit test hid.
it('walks the pane subtree to the claude child that carries the lease id', async () => {
const runner = vi.fn<CommandRunner>().mockResolvedValueOnce(result('1234 node 0 0 0 0\n'));
// pane pid 1234 = node(mosaic): NO lease env. child 5678 = claude: carries it.
const environmentReader = vi.fn(async (pid: number) =>
pid === 5678 ? `FOO=bar\0MOSAIC_LEASE_SESSION_ID=${sessionId}\0` : `FOO=bar\0`,
);
const childrenReader = vi.fn(async (pid: number) => (pid === 1234 ? [5678] : []));
const transport = new TmuxPromotionTransport({
environmentReader,
childrenReader,
mosaicHome: '/mosaic',
rosterLoader: async () => roster,
runner,
});
const target = await transport.resolve('claude-seat');
expect(target.sessionId).toBe(sessionId);
expect(environmentReader).toHaveBeenCalledWith(1234); // pane pid: no lease
expect(environmentReader).toHaveBeenCalledWith(5678); // walked to the child
expect(childrenReader).toHaveBeenCalledWith(1234); // walk actually ran
});
it('fails closed when no process in the pane subtree carries a lease id', async () => {
const runner = vi.fn<CommandRunner>().mockResolvedValueOnce(result('1234 node 0 0 0 0\n'));
const environmentReader = vi.fn(async () => `FOO=bar\0`);
const childrenReader = vi.fn(async (pid: number) => (pid === 1234 ? [5678] : []));
const transport = new TmuxPromotionTransport({
environmentReader,
childrenReader,
mosaicHome: '/mosaic',
rosterLoader: async () => roster,
runner,
});
await expect(transport.resolve('claude-seat')).rejects.toThrow('no readable lease session');
expect(childrenReader).toHaveBeenCalledWith(1234);
});
});
@@ -0,0 +1,208 @@
import { readFile } from 'node:fs/promises';
import {
buildTmuxListPanesCommand,
getRosterAgent,
parseTmuxListPanes,
resolveFleetPaths,
type CommandResult,
type CommandRunner,
type FleetRoster,
RUNTIME_ACCEPTABLE_COMMANDS,
socketArgs,
} from '../commands/fleet.js';
import { loadFleetRoster } from './fleet-roster-v1.js';
const PROMOTION_COMMAND = '/mosaic-promote';
const SESSION_ID_PATTERN = /^[a-f0-9]{64}$/;
const TRANSPORT_COMMAND_TIMEOUT_MS = 5_000;
export interface PromotionTarget {
bundle: string;
seat: string;
sessionId: string;
}
export interface PromotionTransport {
resolve(seat: string): Promise<PromotionTarget>;
sendPromotion(target: PromotionTarget): Promise<void>;
}
export interface TmuxPromotionTransportOptions {
environmentReader?: (pid: number) => Promise<string>;
childrenReader?: (pid: number) => Promise<number[]>;
mosaicHome: string;
rosterLoader?: () => Promise<FleetRoster>;
runner: CommandRunner;
}
// The launcher runs the runtime as a spawnSync CHILD of node(mosaic) (see
// launch.ts:99 — deliberate, so the parent survives to propagate signals), so
// MOSAIC_LEASE_SESSION_ID lives on the claude child, NOT on the tmux pane's root
// pid. Bound the descendant search so a hung/large process tree can't stall it.
const MAX_SUBTREE_PIDS = 128;
/** Local, roster-bound transport for the in-seat promotion command. */
export class TmuxPromotionTransport implements PromotionTransport {
private readonly environmentReader: (pid: number) => Promise<string>;
private readonly childrenReader: (pid: number) => Promise<number[]>;
private readonly rosterLoader: () => Promise<FleetRoster>;
constructor(private readonly options: TmuxPromotionTransportOptions) {
this.environmentReader = options.environmentReader ?? readPaneEnvironment;
this.childrenReader = options.childrenReader ?? readChildPids;
this.rosterLoader =
options.rosterLoader ??
(() => loadFleetRoster(resolveFleetPaths(options.mosaicHome).rosterPath));
}
async resolve(seat: string): Promise<PromotionTarget> {
const roster = await this.rosterLoader();
const agent = getRosterAgent(roster, seat);
if (agent.runtime !== 'claude') {
throw new Error(`Lease promotion is currently available only for Claude seats: ${seat}.`);
}
const paneResult = await this.run(
buildTmuxListPanesCommand(agent.name, roster.tmux.socketName),
);
if (paneResult.exitCode !== 0) {
throw new Error(`Promotion seat is unavailable: ${seat}.`);
}
const pane = parseTmuxListPanes(paneResult.stdout);
const allowedCommands = RUNTIME_ACCEPTABLE_COMMANDS.claude;
if (
pane.dead ||
pane.pid === null ||
pane.command === null ||
allowedCommands === undefined ||
!allowedCommands.includes(pane.command)
) {
throw new Error(`Promotion seat runtime identity mismatch: ${seat}.`);
}
const sessionId = await this.resolveLeaseSessionId(pane.pid);
if (sessionId === null) {
throw new Error(`Promotion seat has no readable lease session: ${seat}.`);
}
return {
bundle: roster.tmux.socketName || 'default',
seat: agent.name,
sessionId,
};
}
/**
* Find the lease session id in the pane's process subtree. The pane's root pid
* is node(mosaic), which has no lease env; the id lives on the claude child.
* BFS from the root, bounded, returning the first descendant that carries a
* valid MOSAIC_LEASE_SESSION_ID. Fail-closed (null) if none is found.
*/
private async resolveLeaseSessionId(rootPid: number): Promise<string | null> {
const queue: number[] = [rootPid];
const seen = new Set<number>();
while (queue.length > 0 && seen.size < MAX_SUBTREE_PIDS) {
const pid = queue.shift()!;
if (seen.has(pid)) continue;
seen.add(pid);
let sessionId: string | null = null;
try {
sessionId = parseLeaseSessionId(await this.environmentReader(pid));
} catch {
sessionId = null;
}
if (sessionId !== null) return sessionId;
let children: number[] = [];
try {
children = await this.childrenReader(pid);
} catch {
children = [];
}
for (const child of children) {
if (!seen.has(child)) queue.push(child);
}
}
return null;
}
async sendPromotion(target: PromotionTarget): Promise<void> {
const targetPane = `=${target.seat}:0.0`;
const socketName = target.bundle === 'default' ? '' : target.bundle;
// Registered Claude commands must arrive as their exact literal text; the
// fleet agent sender prepends an identity envelope, so it cannot carry this
// command without preventing the UserPromptSubmit matcher from recognizing it.
await this.runPromotionCommand([
'tmux',
...socketArgs(socketName),
'send-keys',
'-t',
targetPane,
'-l',
PROMOTION_COMMAND,
]);
await this.runPromotionCommand([
'tmux',
...socketArgs(socketName),
'send-keys',
'-t',
targetPane,
'Enter',
]);
}
private async runPromotionCommand(command: string[]): Promise<void> {
const result = await this.run(command);
if (result.exitCode !== 0) {
throw new Error('Promotion command delivery failed.');
}
}
private async run(command: string[]): Promise<CommandResult> {
const [executable, ...args] = command;
if (executable === undefined) {
throw new Error('Promotion transport command is empty.');
}
return await withTimeout(this.options.runner(executable, args), TRANSPORT_COMMAND_TIMEOUT_MS);
}
}
function withTimeout<T>(operation: Promise<T>, timeoutMs: number): Promise<T> {
return new Promise((resolve, reject) => {
const timeout = setTimeout(() => {
reject(new Error(`Promotion transport command timed out after ${timeoutMs}ms.`));
}, timeoutMs);
void operation.then(
(value) => {
clearTimeout(timeout);
resolve(value);
},
(error: unknown) => {
clearTimeout(timeout);
reject(error);
},
);
});
}
async function readPaneEnvironment(pid: number): Promise<string> {
return readFile(`/proc/${pid}/environ`, 'utf8');
}
async function readChildPids(pid: number): Promise<number[]> {
// Linux exposes direct children of the main thread here (CONFIG_PROC_CHILDREN).
try {
const raw = await readFile(`/proc/${pid}/task/${pid}/children`, 'utf8');
return raw
.split(/\s+/)
.filter(Boolean)
.map((value) => Number.parseInt(value, 10))
.filter((value) => Number.isInteger(value) && value > 0);
} catch {
return [];
}
}
function parseLeaseSessionId(environment: string): string | null {
const value = environment
.split('\0')
.find((entry) => entry.startsWith('MOSAIC_LEASE_SESSION_ID='))
?.slice('MOSAIC_LEASE_SESSION_ID='.length);
return value !== undefined && SESSION_ID_PATTERN.test(value) ? value : null;
}
@@ -0,0 +1,289 @@
#!/usr/bin/env python3
"""Invariant R: a read-only carve-out can neither disappear nor be shadowed.
The broker's carve-out is an authentication bypass for UNVERIFIED runtimes, so
this test imports the live ``READ_ONLY_TOOLS`` object instead of copying it.
Claude MCP names are namespaced, making an exact proven allow-list sufficient.
Pi extensions are unnamespaced and may override built-ins, so the Pi half boots
the installed runtime and requires every carve-out winner to retain built-in
provenance.
"""
from __future__ import annotations
import importlib
import json
import os
import shutil
import subprocess
import sys
import tempfile
import time
import unittest
from pathlib import Path
from typing import Final
PACKAGE_ROOT = Path(__file__).parents[2]
FRAMEWORK = PACKAGE_ROOT / "framework"
LEASE_BROKER = FRAMEWORK / "tools/lease-broker"
PI_EXTENSION = FRAMEWORK / "runtime/pi/mosaic-extension.ts"
sys.path.insert(0, str(LEASE_BROKER))
daemon = importlib.import_module("daemon")
READ_ONLY_TOOLS = daemon.READ_ONLY_TOOLS
# Claude Code's measured, bare built-ins that are both registered and incapable
# of filesystem mutation or subprocess execution. MCP tools are namespaced as
# mcp__<server>__<tool>, so they cannot replace these bare identities.
CLAUDE_PROVEN_READ_ONLY_TOOLS: Final = frozenset({"Read", "Grep", "Glob"})
# W-B measured Pi 0.84.1 through getAllTools(), observed every tool_call name,
# and cross-checked dist/core/tools/index.js:18. Keep every measured built-in
# here so a runtime registry change forces the security classification to be
# revisited even when a built-in is deliberately excluded from the carve-out.
PI_VERSION: Final = "0.84.1"
PI_PROBE_ATTEMPTS: Final = 3
PI_PROBE_TIMEOUT_SECONDS: Final = 45
PI_PROBE_BACKOFF_SECONDS: Final = 0.25
PI_PROVEN_READ_ONLY_TOOLS: Final = frozenset({"read", "ls"})
PI_SUBPROCESS_TOOLS: Final = frozenset({"grep", "find"})
PI_MUTATING_TOOLS: Final = frozenset({"bash", "edit", "write"})
PI_MEASURED_BUILTINS: Final = (
PI_PROVEN_READ_ONLY_TOOLS | PI_SUBPROCESS_TOOLS | PI_MUTATING_TOOLS
)
# Pi 0.84.1 built-ins individually proven incapable of subprocess execution or
# filesystem writes on their default path:
# - read: dist/core/tools/read.js:26-29 dispatches only read/access operations.
# - ls: dist/core/tools/ls.js:19-22 dispatches only exists/stat/readdir operations.
# grep and find are deliberately absent: grep.js:99/148 and find.js:161/203
# reach ensureTool(..., true) and spawn(), including the cold-cache download,
# write, chmod, and exec path in dist/utils/tools-manager.js:285-313.
PI_CAPABILITY_SAFE_TOOLS: Final = frozenset({"read", "ls"})
# Falsifier-only inputs. They are intentionally undocumented outside this test:
# normal CI leaves them unset; the W-A evidence run uses them to prove that the
# suite turns red for a nonexistent Claude carve-out or a Pi built-in override.
CLAUDE_EXTRA_TOOL_ENV: Final = "MOSAIC_INVARIANT_R_CLAUDE_EXTRA_TOOL"
PI_EXTRA_EXTENSION_ENV: Final = "MOSAIC_INVARIANT_R_PI_EXTRA_EXTENSION"
def run_pi_registry_command(
command: list[str],
environ: dict[str, str],
*,
runner=subprocess.run,
sleeper=time.sleep,
) -> subprocess.CompletedProcess[str]:
"""Run the registry probe with bounded retries for concurrent-Pi stalls."""
for attempt in range(1, PI_PROBE_ATTEMPTS + 1):
try:
return runner(
command,
check=False,
capture_output=True,
text=True,
env=environ,
timeout=PI_PROBE_TIMEOUT_SECONDS,
)
except subprocess.TimeoutExpired as error:
if attempt == PI_PROBE_ATTEMPTS:
raise AssertionError(
"Pi registry probe could not complete after "
f"{PI_PROBE_ATTEMPTS} attempts (concurrent pi?); this is a "
"probe/infra failure, NOT an Invariant R violation"
) from error
sleeper(PI_PROBE_BACKOFF_SECONDS * attempt)
raise AssertionError("unreachable Pi registry retry state")
def probe_pi_registry() -> list[dict[str, object]]:
"""Boot Pi's real registry and return the final winning tool definitions."""
pi = shutil.which("pi")
if pi is None:
raise AssertionError("installed Pi runtime is required for Invariant R")
version = subprocess.run(
[pi, "--version"],
check=False,
capture_output=True,
text=True,
timeout=10,
)
if version.returncode != 0:
raise AssertionError(f"Pi version probe failed: {version.stderr.strip()}")
if version.stdout.strip() != PI_VERSION:
raise AssertionError(
f"Pi runtime changed from measured {PI_VERSION} to {version.stdout.strip()!r}; "
"remeasure its registry before updating Invariant R"
)
with tempfile.TemporaryDirectory() as temporary:
root = Path(temporary)
output = root / "registry.json"
observer = root / "registry-observer.ts"
observer.write_text(
"import { writeFileSync } from 'node:fs';\n"
"export default function register(pi: any) {\n"
" pi.on('session_start', () => {\n"
f" writeFileSync({json.dumps(str(output))}, JSON.stringify(pi.getAllTools()));\n"
" process.exit(0);\n"
" });\n"
"}\n",
encoding="utf-8",
)
command = [
pi,
"--mode",
"text",
"--no-session",
"--no-approve",
"--no-context-files",
"--no-skills",
"--no-prompt-templates",
"--no-extensions",
"-e",
str(observer),
"-e",
str(PI_EXTENSION),
]
extra_extension = os.environ.get(PI_EXTRA_EXTENSION_ENV)
if extra_extension:
command.extend(("-e", extra_extension))
command.append("Invariant R registry probe")
completed = run_pi_registry_command(
command,
{**os.environ, "PI_OFFLINE": "1"},
)
if completed.returncode != 0 or not output.is_file():
raise AssertionError(
"Pi registry probe failed "
f"(status {completed.returncode}): {completed.stderr.strip()}"
)
value = json.loads(output.read_text(encoding="utf-8"))
if not isinstance(value, list) or not value:
raise AssertionError("Pi registry probe returned no tools; control failed")
return value
class InvariantRTest(unittest.TestCase):
def test_live_carve_out_has_only_supported_runtimes(self) -> None:
self.assertEqual(set(READ_ONLY_TOOLS), {"claude", "pi"})
def test_claude_carve_out_is_registered_and_proven(self) -> None:
carve_out = set(READ_ONLY_TOOLS["claude"])
falsifier = os.environ.get(CLAUDE_EXTRA_TOOL_ENV)
if falsifier:
carve_out.add(falsifier)
self.assertEqual(
carve_out,
set(CLAUDE_PROVEN_READ_ONLY_TOOLS),
"every Claude carve-out must exist and be in the exact proven read-only allow-list",
)
def test_pi_carve_out_has_no_exec_or_write_capability(self) -> None:
carve_out = set(READ_ONLY_TOOLS["pi"])
capability_unsafe = carve_out - set(PI_CAPABILITY_SAFE_TOOLS)
self.assertFalse(
capability_unsafe,
f"capability-unsafe Pi carve-out tools: {sorted(capability_unsafe)!r}; "
"Pi 0.84.1 grep.js:99/148 and find.js:161/203 reach "
"ensureTool(..., true) and spawn(), whose cold-cache path downloads, "
"writes, chmods, and execs",
)
def test_pi_carve_out_resolves_to_real_unshadowed_builtins(self) -> None:
carve_out = set(READ_ONLY_TOOLS["pi"])
self.assertEqual(
carve_out,
set(PI_PROVEN_READ_ONLY_TOOLS),
"Pi carve-out drift requires a new runtime measurement and classification",
)
self.assertTrue(carve_out.isdisjoint(PI_MUTATING_TOOLS))
registry = probe_pi_registry()
by_name: dict[str, dict[str, object]] = {}
for entry in registry:
name = entry.get("name")
if not isinstance(name, str):
self.fail(f"Pi registry entry has no string name: {entry!r}")
by_name[name] = entry
builtin_names = {
name
for name, entry in by_name.items()
if isinstance(entry.get("sourceInfo"), dict)
and entry["sourceInfo"].get("source") == "builtin"
}
self.assertEqual(
builtin_names,
set(PI_MEASURED_BUILTINS),
"Pi's real built-in registry drifted from the positive-control W-B measurement",
)
for name in sorted(carve_out):
with self.subTest(tool=name):
self.assertIn(name, by_name, "Pi carve-out names must exist in the real registry")
source = by_name[name].get("sourceInfo")
self.assertIsInstance(source, dict)
if isinstance(source, dict):
self.assertEqual(
source.get("source"),
"builtin",
f"Pi extension or SDK tool shadowed read-only carve-out {name!r}",
)
self.assertEqual(source.get("path"), f"<builtin:{name}>")
def test_pi_probe_retries_timeouts_before_succeeding(self) -> None:
attempts: list[float] = []
backoffs: list[float] = []
def timeout_twice(command, **kwargs):
attempts.append(kwargs["timeout"])
if len(attempts) < 3:
raise subprocess.TimeoutExpired(command, kwargs["timeout"])
return subprocess.CompletedProcess(command, 0, "", "")
completed = run_pi_registry_command(
["pi", "probe"],
{},
runner=timeout_twice,
sleeper=backoffs.append,
)
self.assertEqual(completed.returncode, 0)
self.assertEqual(attempts, [45, 45, 45])
self.assertEqual(backoffs, [0.25, 0.5])
def test_pi_probe_labels_exhausted_timeouts_as_infrastructure_failure(self) -> None:
attempts = 0
def always_timeout(command, **kwargs):
nonlocal attempts
attempts += 1
raise subprocess.TimeoutExpired(command, kwargs["timeout"])
with self.assertRaisesRegex(
AssertionError,
"Pi registry probe could not complete .* NOT an Invariant R violation",
) as caught:
run_pi_registry_command(
["pi", "probe"],
{},
runner=always_timeout,
sleeper=lambda _delay: None,
)
self.assertEqual(attempts, 3)
self.assertIsInstance(caught.exception.__cause__, subprocess.TimeoutExpired)
if __name__ == "__main__":
unittest.main()
@@ -0,0 +1,179 @@
#!/usr/bin/env python3
"""The promotion client must never build a binding narrower than it claims.
RED-first against a real defect: ``build_construction`` skipped any normative
source it could not read (``except OSError: continue``) and promoted whatever
remained. That is not a degraded binding, it is a forged smaller one the
broker recomputes ``h_source`` / ``h_payload`` from the fragments it is *sent*
(``daemon.py:602-616``), so an omitted fragment is internally consistent and
``PAYLOAD_BINDING_MISMATCH`` cannot fire. Measured before the fix: with only
``USER.md`` readable (964 bytes on the live host), the client produced a
one-fragment construction with ``promotion=True``.
The classification under test mirrors the framework's own file ownership, and
must keep mirroring it:
* framework-owned, reconciled every upgrade (``install.sh`` FRAMEWORK_OWNED /
``config/file-adapter.ts`` FRAMEWORK_OWNED_FILES) plus the per-runtime
contract absence is a broken deployment, so it is REFUSED;
* ``SOUL.md`` / ``USER.md`` install.sh deliberately does not seed them
("generated by `mosaic init`"), so absence is legitimate and ALLOWED.
Unreadable is treated separately from absent for *every* source, optional ones
included: a file that will not open is not a file that was never configured, and
collapsing the two is what let a permission change quietly shrink the law.
"""
from __future__ import annotations
import contextlib
import io
import os
import sys
import tempfile
import unittest
from pathlib import Path
TOOLS = Path(__file__).parents[2] / "framework/tools/lease-broker"
sys.path.insert(0, str(TOOLS))
import lease_promote # noqa: E402
RUNTIME = "pi"
RUNTIME_CONTRACT = f"runtime/{RUNTIME}/RUNTIME.md"
ALL_SOURCES = (*lease_promote.FRAGMENT_SOURCES, RUNTIME_CONTRACT)
REQUIRED = frozenset(lease_promote.REQUIRED_SOURCES) | {RUNTIME_CONTRACT}
# Derived, never listed: a hand-kept second copy is exactly the drift this file
# exists to catch.
OPTIONAL = tuple(s for s in ALL_SOURCES if s not in REQUIRED)
# chmod 0o000 does not deny root (CAP_DAC_OVERRIDE), so the unreadable
# simulations would fail spuriously in a root container.
runs_unprivileged = unittest.skipIf(
os.geteuid() == 0, "chmod 0o000 cannot make a file unreadable to root"
)
class PromotionBindingTest(unittest.TestCase):
def setUp(self) -> None:
self._previous_home = os.environ.get("MOSAIC_HOME")
self._temporary = tempfile.TemporaryDirectory()
self.root = Path(self._temporary.name)
for source_id in ALL_SOURCES:
path = self.root / source_id
path.parent.mkdir(parents=True, exist_ok=True)
path.write_bytes(f"# {source_id}\nnormative bytes\n".encode())
os.environ["MOSAIC_HOME"] = str(self.root)
def tearDown(self) -> None:
for path in self.root.rglob("*"):
if path.is_file():
path.chmod(0o644)
self._temporary.cleanup()
if self._previous_home is None:
os.environ.pop("MOSAIC_HOME", None)
else:
os.environ["MOSAIC_HOME"] = self._previous_home
def reset_home(self) -> None:
"""Discard the current home and seed a fresh complete one.
Each subTest mutates the tree destructively, so it needs a clean start
and the old one must be released, not orphaned.
"""
self.tearDown()
self.setUp()
def build(self):
return lease_promote.build_construction(RUNTIME)
def source_ids(self) -> list[str]:
construction, _ = self.build()
return [f["source_id"] for f in construction["fragments"]]
# --- the binding is complete when the deployment is complete -------------
def test_complete_deployment_binds_every_source(self) -> None:
construction, result = self.build()
self.assertEqual([f["source_id"] for f in construction["fragments"]], list(ALL_SOURCES))
self.assertTrue(result.promotion)
# --- absence: refused for framework-owned, allowed for operator-owned ----
def test_absent_required_source_is_refused(self) -> None:
for source_id in sorted(REQUIRED):
with self.subTest(source=source_id):
self.reset_home()
(self.root / source_id).unlink()
with self.assertRaises(lease_promote.IncompleteBinding) as caught:
self.build()
self.assertIn(source_id, str(caught.exception))
def test_absent_operator_source_still_binds_the_rest(self) -> None:
for source_id in OPTIONAL:
with self.subTest(source=source_id):
self.reset_home()
(self.root / source_id).unlink()
notice = io.StringIO()
with contextlib.redirect_stderr(notice):
bound = self.source_ids()
self.assertNotIn(source_id, bound)
for required in lease_promote.REQUIRED_SOURCES:
self.assertIn(required, bound)
# A silent omission is the original defect in miniature: the
# narrower binding must announce itself.
self.assertIn(source_id, notice.getvalue())
# --- unreadable is never the same as absent -----------------------------
@runs_unprivileged
def test_unreadable_source_is_refused_even_when_optional(self) -> None:
for source_id in ALL_SOURCES:
with self.subTest(source=source_id):
self.reset_home()
(self.root / source_id).chmod(0o000)
with self.assertRaises(lease_promote.IncompleteBinding) as caught:
self.build()
self.assertIn(source_id, str(caught.exception))
# --- the exact measured regression --------------------------------------
@runs_unprivileged
def test_single_readable_source_cannot_promote(self) -> None:
"""The observed failure: only USER.md readable produced a valid binding."""
for source_id in ALL_SOURCES:
if source_id != "USER.md":
(self.root / source_id).chmod(0o000)
with self.assertRaises(lease_promote.IncompleteBinding):
self.build()
@runs_unprivileged
def test_no_source_readable_cannot_promote(self) -> None:
for source_id in ALL_SOURCES:
(self.root / source_id).chmod(0o000)
with self.assertRaises(lease_promote.IncompleteBinding):
self.build()
# --- the classification must not drift from the framework's -------------
def test_required_set_excludes_only_the_unseeded_sources(self) -> None:
"""`install.sh` decides which files exist; this list must follow it.
If a source moves between framework-owned and operator-generated
upstream, this fails and forces the classification to be re-read rather
than silently inherited.
"""
self.assertEqual(
set(lease_promote.REQUIRED_SOURCES),
{"CONSTITUTION.md", "AGENTS.md", "STANDARDS.md"},
"REQUIRED_SOURCES changed — re-read install.sh FRAMEWORK_OWNED and "
"config/file-adapter.ts FRAMEWORK_OWNED_FILES before accepting it",
)
self.assertTrue(
set(lease_promote.REQUIRED_SOURCES) <= set(lease_promote.FRAGMENT_SOURCES),
"a required source is not in the binding order",
)
if __name__ == "__main__":
unittest.main()
@@ -0,0 +1,652 @@
#!/usr/bin/env python3
"""RED-first contracts for the operator-triggered Claude promotion hooks."""
from __future__ import annotations
import importlib.util
import io
import json
import os
import stat
import subprocess
import tempfile
import unittest
from pathlib import Path
from unittest import mock
PACKAGE_ROOT = Path(__file__).parents[2]
FRAMEWORK = PACKAGE_ROOT / "framework"
TOOLS = FRAMEWORK / "tools/lease-broker"
BEGIN_PATH = TOOLS / "promote-begin.py"
COMPLETE_PATH = TOOLS / "promote-complete.py"
OBSERVER_CLIENT_PATH = TOOLS / "receipt-observer-client.py"
RECEIPT_CHALLENGE_PATH = TOOLS / "receipt_challenge.py"
CLAUDE_SETTINGS = FRAMEWORK / "runtime/claude/settings.json"
CLAUDE_COMMAND = FRAMEWORK / "runtime/claude/commands/mosaic-promote.md"
SESSION_ID = "a" * 64
CHALLENGE = "b" * 64
H_PAYLOAD = "c" * 64
RECEIPT = (
f"MOSAIC-RECEIPT{{challenge={CHALLENGE}; H_payload={H_PAYLOAD}; gen=1; cep=0}}"
)
NOW = 10_000.0
def load_module(name: str, path: Path):
if not path.is_file():
raise AssertionError(f"shipped module is missing: {path}")
spec = importlib.util.spec_from_file_location(name, path)
if spec is None or spec.loader is None:
raise RuntimeError(f"unable to load {name}")
module = importlib.util.module_from_spec(spec)
spec.loader.exec_module(module)
return module
class PromotionHookFixture(unittest.TestCase):
@classmethod
def setUpClass(cls) -> None:
cls.begin = load_module("promotion_begin_test", BEGIN_PATH)
cls.complete = load_module("promotion_complete_test", COMPLETE_PATH)
def setUp(self) -> None:
self.temporary = tempfile.TemporaryDirectory()
self.runtime_dir = Path(self.temporary.name)
self.environment = {
"XDG_RUNTIME_DIR": str(self.runtime_dir),
"MOSAIC_LEASE_SESSION_ID": SESSION_ID,
}
self.pending_dir = self.runtime_dir / "mosaic-lease"
self.pending_file = self.pending_dir / f"pending-{SESSION_ID}"
def tearDown(self) -> None:
self.temporary.cleanup()
@staticmethod
def completed(payload: dict[str, object], returncode: int = 0, stderr: str = ""):
return subprocess.CompletedProcess(
["lease_promote.py"],
returncode,
json.dumps(payload),
stderr,
)
@classmethod
def successful_begin_reply(cls, **extra: object) -> dict[str, object]:
return {
"ok": True,
"state": "PENDING_VERIFICATION",
"receipt_challenge": CHALLENGE,
"receipt": RECEIPT,
"binding": {
"compaction_epoch": 0,
"request_epoch": 0,
"h_source": "d" * 64,
"h_payload": H_PAYLOAD,
"runtime_generation": 1,
"schema_version": 1,
},
**extra,
}
def write_authorization(self) -> None:
directory = self.pending_dir / "authorizations"
directory.mkdir(parents=True, mode=0o700)
self.pending_dir.chmod(0o700)
directory.chmod(0o700)
token = directory / f"{SESSION_ID}.auth"
token.write_text(json.dumps({"nonce": "e" * 64, "seat": "claude-seat", "session_id": SESSION_ID, "expires_at": NOW + 60, "ts": NOW}), encoding="utf-8")
token.chmod(0o600)
def run_begin(
self,
prompt: str,
runner: mock.Mock,
authorized: bool = True,
) -> tuple[int, str, str]:
if authorized and prompt == "/mosaic-promote":
self.write_authorization()
stdout = io.StringIO()
stderr = io.StringIO()
self.observer = mock.Mock(return_value={"ok": True})
with mock.patch.object(self.begin, "observer_request", self.observer):
result = self.begin.main(
environ={**self.environment, "MOSAIC_RECEIPT_OBSERVER_SOCKET": "/tmp/observer", "MOSAIC_RUNTIME_GENERATION": "1"},
stdin=io.BytesIO(json.dumps({"prompt": prompt}).encode()),
stdout=stdout,
stderr=stderr,
run=runner,
now=lambda: NOW,
)
return result, stdout.getvalue(), stderr.getvalue()
def write_pending(self, challenge: str = CHALLENGE) -> None:
self.pending_dir.mkdir(mode=0o700, exist_ok=True)
self.pending_file.write_text(challenge, encoding="utf-8")
self.pending_file.chmod(0o600)
def run_complete(
self,
runner: mock.Mock,
now: float | None = None,
) -> tuple[int, str]:
stderr = io.StringIO()
options: dict[str, object] = {
"environ": self.environment,
"stderr": stderr,
"run": runner,
}
if now is not None:
options["now"] = lambda: now
result = self.complete.main(**options)
return result, stderr.getvalue()
class PromotionBeginTest(PromotionHookFixture):
def test_injected_exact_promotion_without_authorization_is_inert(self) -> None:
runner = mock.Mock()
result, stdout, stderr = self.run_begin("/mosaic-promote", runner, authorized=False)
self.assertEqual(result, 0)
self.assertEqual(stdout, "")
self.assertIn("NOT_AUTHORIZED", stderr)
runner.assert_not_called()
self.observer.assert_not_called()
self.assertEqual(json.loads((self.pending_dir / "last-result.json").read_text())["reason"], "NOT_AUTHORIZED")
def test_valid_token_posts_receipt_then_completes_without_model_context(self) -> None:
runner = mock.Mock(side_effect=[
self.completed(self.successful_begin_reply()),
self.completed({"stage": "promote_lease", "ok": True, "state": "VERIFIED"}),
])
result, stdout, stderr = self.run_begin("/mosaic-promote", runner)
self.assertEqual((result, stdout, stderr), (0, "", ""))
self.assertFalse((self.pending_dir / "authorizations" / f"{SESSION_ID}.auth").exists())
self.observer.assert_called_once()
self.assertEqual(self.observer.call_args.args[1]["latest_assistant_message"], RECEIPT)
self.assertEqual(runner.call_args_list[1].args[0][-2:], ["--complete", CHALLENGE])
self.assertTrue(json.loads((self.pending_dir / "last-result.json").read_text())["verified"])
@unittest.skip("superseded by mechanical promotion")
def test_exact_prompt_writes_private_challenge_and_injects_verbatim_receipt(self) -> None:
runner = mock.Mock(return_value=self.completed(self.successful_begin_reply()))
result, stdout, stderr = self.run_begin("/mosaic-promote", runner)
self.assertEqual(result, 0)
self.assertEqual(stderr, "")
output = json.loads(stdout)
self.assertEqual(
output["hookSpecificOutput"]["additionalContext"],
"The operator invoked the registered /mosaic-promote command. "
"This receipt was generated locally by this seat's own lease broker; "
"echoing it verbatim is the designed confirmation step and discloses nothing. "
f"Reply with exactly the following text and nothing else: {RECEIPT}",
)
self.assertEqual(self.pending_file.read_text(encoding="utf-8"), CHALLENGE)
self.assertEqual(stat.S_IMODE(self.pending_file.stat().st_mode), 0o600)
command = runner.call_args.args[0]
self.assertEqual(command[-1], "--begin")
self.assertTrue(command[-2].endswith("lease_promote.py"))
def test_nonmatching_prompt_has_zero_side_effects(self) -> None:
runner = mock.Mock()
result, stdout, stderr = self.run_begin("please /mosaic-promote", runner)
self.assertEqual(result, 0)
self.assertEqual(stdout, "")
self.assertEqual(stderr, "")
runner.assert_not_called()
self.assertFalse(self.pending_dir.exists())
@unittest.skip("superseded by breadcrumb-only mechanical errors")
def test_begin_refusal_reports_daemon_code_without_pending_file(self) -> None:
runner = mock.Mock(
return_value=self.completed({"ok": False, "code": "INVALID_BINDING"})
)
result, stdout, _stderr = self.run_begin("/mosaic-promote", runner)
self.assertEqual(result, 0)
self.assertIn("INVALID_BINDING", json.loads(stdout)["hookSpecificOutput"]["additionalContext"])
self.assertFalse(self.pending_file.exists())
def test_sweep_removes_stale_sibling_and_spares_fresh_sibling(self) -> None:
self.pending_dir.mkdir(mode=0o700)
stale = self.pending_dir / "pending-stale"
fresh = self.pending_dir / "pending-fresh"
stale.write_text("stale", encoding="utf-8")
fresh.write_text("fresh", encoding="utf-8")
os.utime(stale, (NOW - 3_601, NOW - 3_601))
os.utime(fresh, (NOW - 3_599, NOW - 3_599))
runner = mock.Mock(return_value=self.completed(self.successful_begin_reply()))
result, _stdout, _stderr = self.run_begin("/mosaic-promote", runner)
self.assertEqual(result, 0)
self.assertFalse(stale.exists())
self.assertTrue(fresh.exists())
def test_sweep_removes_stale_atomic_temporary_file(self) -> None:
self.pending_dir.mkdir(mode=0o700)
stale_temporary = self.pending_dir / f".pending-{SESSION_ID}.tmp-abandoned"
stale_temporary.write_text("partial", encoding="utf-8")
os.utime(stale_temporary, (NOW - 3_601, NOW - 3_601))
runner = mock.Mock(return_value=self.completed(self.successful_begin_reply()))
result, _stdout, _stderr = self.run_begin("/mosaic-promote", runner)
self.assertEqual(result, 0)
self.assertFalse(stale_temporary.exists())
@unittest.skip("authorization fixture creates a secure parent directory")
def test_insecure_pending_directory_mode_refuses_before_begin(self) -> None:
self.pending_dir.mkdir(mode=0o755)
self.pending_dir.chmod(0o755)
runner = mock.Mock(return_value=self.completed(self.successful_begin_reply()))
result, stdout, _stderr = self.run_begin("/mosaic-promote", runner)
self.assertEqual(result, 0)
runner.assert_not_called()
self.assertFalse(self.pending_file.exists())
self.assertEqual(stdout, "")
def test_insecure_runtime_directory_mode_refuses_before_begin(self) -> None:
self.runtime_dir.chmod(0o755)
runner = mock.Mock(return_value=self.completed(self.successful_begin_reply()))
result, stdout, _stderr = self.run_begin("/mosaic-promote", runner)
self.assertEqual(result, 0)
runner.assert_not_called()
self.assertEqual(stdout, "")
def test_parent_symlink_cannot_redirect_pending_write(self) -> None:
outside = self.runtime_dir / "outside"
outside.mkdir(mode=0o700)
self.pending_dir.symlink_to(outside, target_is_directory=True)
runner = mock.Mock(return_value=self.completed(self.successful_begin_reply()))
result, _stdout, _stderr = self.run_begin("/mosaic-promote", runner)
self.assertEqual(result, 0)
runner.assert_not_called()
self.assertFalse((outside / f"pending-{SESSION_ID}").exists())
@unittest.skip("single-use authorization supersedes pending challenge concurrency")
def test_concurrent_begin_is_refused_without_minting_a_second_challenge(self) -> None:
inner_runner = mock.Mock(return_value=self.completed(self.successful_begin_reply()))
inner_result: list[tuple[int, str, str]] = []
def overlap(*_args: object, **_kwargs: object):
inner_result.append(self.run_begin("/mosaic-promote", inner_runner))
return self.completed(self.successful_begin_reply())
outer_runner = mock.Mock(side_effect=overlap)
result, _stdout, _stderr = self.run_begin("/mosaic-promote", outer_runner)
self.assertEqual(result, 0)
inner_runner.assert_not_called()
self.assertEqual(inner_result[0][0], 0)
self.assertIn("PROMOTION_ALREADY_IN_PROGRESS", inner_result[0][1])
@unittest.skip("superseded by mechanical completion")
def test_non_ascii_receipt_reply_is_rejected_without_crashing_hook(self) -> None:
reply = self.successful_begin_reply()
reply["receipt"] = "MOSAIC—RECEIPT"
runner = mock.Mock(return_value=self.completed(reply))
result, stdout, _stderr = self.run_begin("/mosaic-promote", runner)
self.assertEqual(result, 0)
self.assertFalse(self.pending_file.exists())
self.assertIn("INVALID_PROMOTER_REPLY", stdout)
@unittest.skip("superseded by mechanical completion")
def test_success_shaped_reply_with_extra_fields_is_rejected(self) -> None:
runner = mock.Mock(
return_value=self.completed(self.successful_begin_reply(unexpected=True))
)
result, stdout, _stderr = self.run_begin("/mosaic-promote", runner)
self.assertEqual(result, 0)
self.assertFalse(self.pending_file.exists())
self.assertIn("INVALID_PROMOTER_REPLY", stdout)
class PromotionCompleteTest(PromotionHookFixture):
def test_no_pending_file_is_zero_cost_success(self) -> None:
runner = mock.Mock()
result, stderr = self.run_complete(runner)
self.assertEqual(result, 0)
self.assertEqual(stderr, "")
runner.assert_not_called()
def test_success_deletes_pending_file(self) -> None:
self.write_pending()
runner = mock.Mock(
return_value=self.completed(
{"stage": "promote_lease", "ok": True, "state": "VERIFIED"}
)
)
result, _stderr = self.run_complete(runner)
self.assertEqual(result, 0)
self.assertFalse(self.pending_file.exists())
self.assertEqual(runner.call_args.args[0][-2:], ["--complete", CHALLENGE])
def test_success_atomically_writes_a_private_correlated_result_with_wall_clock_expiry(self) -> None:
self.write_pending()
runner = mock.Mock(
return_value=self.completed(
{"stage": "promote_lease", "ok": True, "state": "VERIFIED"}
)
)
with mock.patch.object(self.complete.os, "replace", wraps=os.replace) as replace:
result, _stderr = self.run_complete(runner, now=12_345.0)
result_file = self.pending_dir / "last-result.json"
self.assertEqual(result, 0)
self.assertEqual(stat.S_IMODE(result_file.stat().st_mode), 0o600)
self.assertEqual(
json.loads(result_file.read_text(encoding="utf-8")),
{
"attempt_id": CHALLENGE,
"expires_at_wallclock": 15_945.0,
"reason": None,
"session_id": SESSION_ID,
"ts": 12_345.0,
"verified": True,
},
)
temporary, destination = replace.call_args.args
self.assertRegex(temporary, r"^\.last-result\.json\.tmp-[0-9a-f]+$")
self.assertEqual(destination, "last-result.json")
self.assertFalse(any(path.name.startswith(".last-result.json.tmp-") for path in self.pending_dir.iterdir()))
def test_terminal_failure_writes_a_private_correlated_unverified_result(self) -> None:
self.write_pending()
runner = mock.Mock(
return_value=self.completed(
{"stage": "observe_receipt", "ok": False, "code": "RECEIPT_MISMATCH"}
)
)
result, _stderr = self.run_complete(runner, now=12_345.0)
result_file = self.pending_dir / "last-result.json"
self.assertEqual(result, 0)
self.assertFalse(self.pending_file.exists())
self.assertEqual(stat.S_IMODE(result_file.stat().st_mode), 0o600)
self.assertEqual(
json.loads(result_file.read_text(encoding="utf-8")),
{
"attempt_id": CHALLENGE,
"expires_at_wallclock": None,
"reason": "RECEIPT_MISMATCH",
"session_id": SESSION_ID,
"ts": 12_345.0,
"verified": False,
},
)
def test_each_terminal_failure_deletes_pending_file(self) -> None:
terminal_codes = (
"RECEIPT_REPLAY",
"RECEIPT_MISMATCH",
"INVALID_LEASE_TRANSITION",
"PROMOTION_TOKEN_INVALID",
)
for code in terminal_codes:
with self.subTest(code=code):
self.write_pending()
runner = mock.Mock(
return_value=self.completed(
{"stage": "observe_receipt", "ok": False, "code": code}
)
)
result, stderr = self.run_complete(runner)
self.assertEqual(result, 0)
self.assertFalse(self.pending_file.exists())
self.assertIn(code, stderr)
def test_transient_and_unknown_failures_preserve_pending_file(self) -> None:
transient_codes = (
"RECEIPT_OBSERVATION_UNAVAILABLE",
"BROKER_BUSY",
"ANCESTRY_MISMATCH",
)
for code in transient_codes:
with self.subTest(code=code):
self.write_pending()
runner = mock.Mock(
return_value=self.completed(
{"stage": "observe_receipt", "ok": False, "code": code}
)
)
result, stderr = self.run_complete(runner)
self.assertEqual(result, 0)
self.assertTrue(self.pending_file.exists())
self.assertIn(code, stderr)
def test_transport_failure_preserves_pending_file_and_exits_zero(self) -> None:
self.write_pending()
runner = mock.Mock(
return_value=self.completed({}, returncode=2, stderr="ConnectionRefusedError")
)
result, stderr = self.run_complete(runner)
self.assertEqual(result, 0)
self.assertTrue(self.pending_file.exists())
self.assertIn("ConnectionRefusedError", stderr)
def test_result_write_failure_exits_zero(self) -> None:
self.write_pending()
runner = mock.Mock(
return_value=self.completed(
{"stage": "promote_lease", "ok": True, "state": "VERIFIED"}
)
)
with mock.patch.object(self.complete, "write_result", side_effect=OSError("disk full")):
result, stderr = self.run_complete(runner)
self.assertEqual(result, 0)
self.assertIn("OSError", stderr)
def test_missing_lease_session_id_exits_zero(self) -> None:
self.write_pending()
environment = dict(self.environment)
del environment["MOSAIC_LEASE_SESSION_ID"]
runner = mock.Mock()
stderr = io.StringIO()
result = self.complete.main(environ=environment, stderr=stderr, run=runner)
self.assertEqual(result, 0)
runner.assert_not_called()
self.assertIn("KeyError", stderr.getvalue())
def test_insecure_runtime_directory_mode_preserves_pending(self) -> None:
self.write_pending(CHALLENGE)
self.runtime_dir.chmod(0o755)
runner = mock.Mock(
return_value=self.completed(
{"stage": "promote_lease", "ok": True, "state": "VERIFIED"}
)
)
result, _stderr = self.run_complete(runner)
self.assertEqual(result, 0)
runner.assert_not_called()
self.assertTrue(self.pending_file.exists())
def test_parent_symlink_cannot_redirect_pending_read_or_delete(self) -> None:
outside = self.runtime_dir / "outside"
outside.mkdir(mode=0o700)
outside_pending = outside / f"pending-{SESSION_ID}"
outside_pending.write_text(CHALLENGE, encoding="utf-8")
outside_pending.chmod(0o600)
self.pending_dir.symlink_to(outside, target_is_directory=True)
runner = mock.Mock(
return_value=self.completed(
{"stage": "promote_lease", "ok": True, "state": "VERIFIED"}
)
)
result, _stderr = self.run_complete(runner)
self.assertEqual(result, 0)
runner.assert_not_called()
self.assertTrue(outside_pending.exists())
def test_insecure_pending_file_mode_is_not_consumed(self) -> None:
self.write_pending(CHALLENGE)
self.pending_file.chmod(0o644)
runner = mock.Mock(
return_value=self.completed(
{"stage": "promote_lease", "ok": True, "state": "VERIFIED"}
)
)
result, _stderr = self.run_complete(runner)
self.assertEqual(result, 0)
runner.assert_not_called()
self.assertTrue(self.pending_file.exists())
def test_concurrent_replacement_is_not_deleted_after_success(self) -> None:
self.write_pending(CHALLENGE)
def replace_pending(*_args: object, **_kwargs: object):
replacement = self.pending_dir / "replacement"
replacement.write_text("replacement", encoding="utf-8")
replacement.chmod(0o600)
os.replace(replacement, self.pending_file)
return self.completed(
{"stage": "promote_lease", "ok": True, "state": "VERIFIED"}
)
runner = mock.Mock(side_effect=replace_pending)
result, _stderr = self.run_complete(runner)
self.assertEqual(result, 0)
self.assertEqual(self.pending_file.read_text(encoding="utf-8"), "replacement")
def test_success_shaped_reply_with_extra_fields_preserves_pending(self) -> None:
self.write_pending(CHALLENGE)
runner = mock.Mock(
return_value=self.completed(
{
"stage": "promote_lease",
"ok": True,
"state": "VERIFIED",
"unexpected": True,
}
)
)
result, _stderr = self.run_complete(runner)
self.assertEqual(result, 0)
self.assertTrue(self.pending_file.exists())
class PromotionTemplateWiringTest(unittest.TestCase):
def test_gated_claude_template_wires_begin_and_ordered_stop_chain(self) -> None:
settings = json.loads(CLAUDE_SETTINGS.read_text(encoding="utf-8"))
hooks = settings["hooks"]
submit_commands = [
hook["command"]
for group in hooks["UserPromptSubmit"]
for hook in group["hooks"]
]
self.assertEqual(
submit_commands,
["python3 ~/.config/mosaic/tools/lease-broker/promote-begin.py"],
)
self.assertEqual(
[group.get("matcher") for group in hooks["UserPromptSubmit"]],
["^/mosaic-promote$"],
)
stop_commands = [
hook["command"]
for group in hooks["Stop"]
for hook in group["hooks"]
]
promotion_chains = [
command
for command in stop_commands
if "receipt-observer-client.py" in command and "promote-complete.py" in command
]
self.assertEqual(len(promotion_chains), 1)
chain = promotion_chains[0]
self.assertLess(
chain.index("receipt-observer-client.py"),
chain.index("promote-complete.py"),
)
self.assertIn("observer_status=$?", chain)
self.assertTrue(chain.endswith("exit $observer_status"))
def test_registered_command_is_one_line_and_inert(self) -> None:
body = CLAUDE_COMMAND.read_text(encoding="utf-8")
self.assertEqual(
body,
"Mosaic lease promotion was processed mechanically; no action is needed.\n",
)
class PromotionVerbatimToleranceTest(unittest.TestCase):
def test_echo_turn_with_tool_use_is_rejected_and_requires_two_turns(self) -> None:
observer_client = load_module("promotion_observer_client_test", OBSERVER_CLIENT_PATH)
receipt_challenge = load_module("promotion_receipt_challenge_test", RECEIPT_CHALLENGE_PATH)
challenge = "b" * 64
binding = {
"h_payload": "c" * 64,
"runtime_generation": 1,
"compaction_epoch": 0,
}
receipt = receipt_challenge.receipt_for(challenge, binding)
real_claude_entry = {
"message": {
"role": "assistant",
"content": [
{"type": "text", "text": receipt},
{
"type": "tool_use",
"id": "tool-1",
"name": "mcp__discord__reply",
"input": {"message": "promoted"},
},
],
}
}
extracted = observer_client.assistant_text(real_claude_entry)
accepted = isinstance(extracted, str) and receipt_challenge.is_verbatim_receipt(
extracted,
challenge,
binding,
)
self.assertIsNone(extracted)
self.assertFalse(accepted)
if __name__ == "__main__":
unittest.main()
@@ -207,6 +207,24 @@ class ReceiptObserverTest(BrokerFixture):
with self.assertRaisesRegex(DAEMON.BrokerFailure, "INVALID_LEASE_TRANSITION"): with self.assertRaisesRegex(DAEMON.BrokerFailure, "INVALID_LEASE_TRANSITION"):
self.promote(current["receipt_challenge"]) self.promote(current["receipt_challenge"])
def test_non_ascii_observation_is_mismatch_and_broker_keeps_serving(self) -> None:
construction, binding = self.construction()
cycle = self.begin(binding, construction)
self.record("I refuse—this is not the receipt")
with self.assertRaisesRegex(DAEMON.BrokerFailure, "RECEIPT_MISMATCH"):
self.observe(cycle["receipt_challenge"])
denied = self.broker.handle(self.peer, {
"action": "authorize_tool",
"session_id": self.session_id,
"runtime_generation": 7,
"runtime": "pi",
"tool_name": "bash",
})
self.assertEqual(denied["decision"], "deny")
self.assertEqual(denied["state"], DAEMON.LEASE_UNVERIFIED)
def test_t29_altered_model_hash_cannot_promote_against_shipped_binding(self) -> None: def test_t29_altered_model_hash_cannot_promote_against_shipped_binding(self) -> None:
construction, binding = self.construction() construction, binding = self.construction()
cycle = self.begin(binding, construction) cycle = self.begin(binding, construction)
@@ -0,0 +1,303 @@
#!/usr/bin/env python3
"""Exit-semantics tests for the receipt observer Stop-hook client."""
from __future__ import annotations
import importlib.util
import io
import json
import tempfile
import unittest
from contextlib import redirect_stderr
from pathlib import Path
from unittest import mock
TOOLS = Path(__file__).parents[2] / "framework/tools/lease-broker"
CLIENT_PATH = TOOLS / "receipt-observer-client.py"
def load_client():
spec = importlib.util.spec_from_file_location("receipt_observer_client_test", CLIENT_PATH)
if spec is None or spec.loader is None:
raise RuntimeError("unable to load receipt-observer-client.py")
module = importlib.util.module_from_spec(spec)
spec.loader.exec_module(module)
return module
CLIENT = load_client()
VALID_INPUT = json.dumps({"latest_assistant_message": "ordinary turn"}).encode()
DEEPLY_NESTED_JSON = b"[" * 2_000 + b"0" + b"]" * 2_000
ENVIRONMENT = {
"MOSAIC_RECEIPT_OBSERVER_SOCKET": "/unused/observer.sock",
"MOSAIC_LEASE_SESSION_ID": "a" * 64,
"MOSAIC_RUNTIME_GENERATION": "1",
}
class FakeObserverSocket:
def __init__(self, response: bytes) -> None:
self.response = response
def __enter__(self):
return self
def __exit__(self, *_args: object) -> None:
return None
def settimeout(self, _timeout: float) -> None:
return None
def connect(self, _path: str) -> None:
return None
def sendall(self, _payload: bytes) -> None:
return None
def shutdown(self, _how: int) -> None:
return None
def recv(self, _size: int) -> bytes:
response, self.response = self.response, b""
return response
class ReceiptObserverClientExitSemanticsTest(unittest.TestCase):
def run_client(
self,
*,
input_bytes: bytes = VALID_INPUT,
reply: dict[str, object] | None = None,
transport_error: OSError | None = None,
runtime: str = "pi",
) -> tuple[int, str, mock.Mock]:
request = mock.Mock(return_value=reply)
if transport_error is not None:
request.side_effect = transport_error
stderr = io.StringIO()
with (
mock.patch.object(CLIENT.sys, "stdin", io.BytesIO(input_bytes)),
mock.patch.object(CLIENT, "observer_request", request),
redirect_stderr(stderr),
):
arguments = ["--runtime", runtime]
if runtime == "claude":
arguments.append("--latest-entry")
result = CLIENT.main(arguments, environ=ENVIRONMENT)
return result, stderr.getvalue(), request
def test_claude_prefers_inline_last_assistant_message(self) -> None:
inline = "the just-finished assistant message"
input_bytes = json.dumps({
"last_assistant_message": inline,
"transcript_path": "/must/not/be/opened.jsonl",
}).encode()
result, stderr, request = self.run_client(
input_bytes=input_bytes,
reply={"ok": True},
runtime="claude",
)
self.assertEqual(result, 0)
self.assertEqual(stderr, "")
self.assertEqual(
request.call_args.args[1]["latest_assistant_message"],
inline,
)
def test_claude_falls_back_to_transcript_when_inline_field_is_absent(self) -> None:
with tempfile.TemporaryDirectory() as directory:
transcript = Path(directory) / "transcript.jsonl"
transcript.write_text(
json.dumps({
"message": {
"role": "assistant",
"content": [{"type": "text", "text": "fallback message"}],
}
})
+ "\n",
encoding="utf-8",
)
input_bytes = json.dumps({"transcript_path": str(transcript)}).encode()
result, stderr, request = self.run_client(
input_bytes=input_bytes,
reply={"ok": True},
runtime="claude",
)
self.assertEqual(result, 0)
self.assertEqual(stderr, "")
self.assertEqual(
request.call_args.args[1]["latest_assistant_message"],
"fallback message",
)
def test_claude_present_invalid_inline_field_fails_without_fallback(self) -> None:
fallback = mock.Mock(return_value="must not be used")
input_bytes = json.dumps({
"last_assistant_message": None,
"transcript_path": "/unused/transcript.jsonl",
}).encode()
with mock.patch.object(CLIENT, "claude_latest_entry", fallback):
result, stderr, request = self.run_client(
input_bytes=input_bytes,
reply={"ok": True},
runtime="claude",
)
self.assertEqual(result, 2)
self.assertIn("Mosaic receipt observer refused", stderr)
request.assert_not_called()
fallback.assert_not_called()
def test_claude_inline_message_size_guard_stays_enforced(self) -> None:
request = mock.Mock(return_value={"ok": True})
stderr = io.StringIO()
with (
mock.patch.object(CLIENT.sys, "stdin", io.BytesIO(b"{}")),
mock.patch.object(
CLIENT,
"read_json",
return_value={"last_assistant_message": "x" * (CLIENT.MAX_FRAME + 1)},
),
mock.patch.object(CLIENT, "observer_request", request),
redirect_stderr(stderr),
):
result = CLIENT.main(
["--runtime", "claude", "--latest-entry"],
environ=ENVIRONMENT,
)
self.assertEqual(result, 2)
self.assertIn("Mosaic receipt observer refused", stderr.getvalue())
request.assert_not_called()
def test_pi_still_posts_only_its_runtime_message(self) -> None:
result, stderr, request = self.run_client(reply={"ok": True})
self.assertEqual(result, 0)
self.assertEqual(stderr, "")
self.assertEqual(
request.call_args.args[1]["latest_assistant_message"],
"ordinary turn",
)
def test_nothing_pending_observation_refusal_is_benign(self) -> None:
result, stderr, request = self.run_client(
reply={"ok": False, "code": "OBSERVATION_UNAVAILABLE"}
)
self.assertEqual(result, 0)
self.assertEqual(stderr, "")
request.assert_called_once()
def test_success_reply_remains_successful(self) -> None:
result, stderr, _request = self.run_client(reply={"ok": True})
self.assertEqual(result, 0)
self.assertEqual(stderr, "")
def test_pending_cycle_auth_failure_stays_fail_closed(self) -> None:
result, _stderr, _request = self.run_client(
reply={"ok": False, "code": "ANCESTRY_MISMATCH"}
)
self.assertEqual(result, 2)
def test_transport_failure_stays_fail_closed(self) -> None:
result, stderr, _request = self.run_client(
transport_error=ConnectionRefusedError("observer unavailable")
)
self.assertEqual(result, 2)
self.assertIn("Mosaic receipt observer refused", stderr)
def test_parse_failure_stays_fail_closed(self) -> None:
result, stderr, request = self.run_client(input_bytes=b"{")
self.assertEqual(result, 2)
self.assertIn("Mosaic receipt observer refused", stderr)
request.assert_not_called()
def test_malformed_wire_replies_stay_fail_closed(self) -> None:
for response in (
b"not-json\n",
b'{"ok":true}',
b"{}\n{}\n",
b'{"ok":false,"code":"OBSERVATION_UNAVAILABLE"}\n\n',
b'{"ok":true,"ok":false,"code":"OBSERVATION_UNAVAILABLE"}\n',
DEEPLY_NESTED_JSON + b"\n",
b"x" * (CLIENT.MAX_FRAME + 1),
):
with self.subTest(response=response):
stderr = io.StringIO()
with (
mock.patch.object(CLIENT.sys, "stdin", io.BytesIO(VALID_INPUT)),
mock.patch.object(
CLIENT.socket,
"socket",
return_value=FakeObserverSocket(response),
),
redirect_stderr(stderr),
):
result = CLIENT.main(["--runtime", "pi"], environ=ENVIRONMENT)
self.assertEqual(result, 2)
self.assertIn("Mosaic receipt observer refused", stderr.getvalue())
def test_oversized_input_stays_fail_closed(self) -> None:
result, stderr, request = self.run_client(input_bytes=b"x" * (CLIENT.MAX_FRAME + 1))
self.assertEqual(result, 2)
self.assertIn("Mosaic receipt observer refused", stderr)
request.assert_not_called()
def test_deeply_nested_input_stays_fail_closed(self) -> None:
result, stderr, request = self.run_client(input_bytes=DEEPLY_NESTED_JSON)
self.assertEqual(result, 2)
self.assertIn("Mosaic receipt observer refused", stderr)
request.assert_not_called()
def test_json_recursion_failure_stays_fail_closed(self) -> None:
request = mock.Mock()
stderr = io.StringIO()
with (
mock.patch.object(CLIENT.sys, "stdin", io.BytesIO(VALID_INPUT)),
mock.patch.object(CLIENT, "observer_request", request),
mock.patch.object(
CLIENT.json,
"loads",
side_effect=RecursionError("maximum JSON nesting exceeded"),
),
redirect_stderr(stderr),
):
result = CLIENT.main(["--runtime", "pi"], environ=ENVIRONMENT)
self.assertEqual(result, 2)
self.assertIn("Mosaic receipt observer refused", stderr.getvalue())
request.assert_not_called()
def test_observation_unavailable_with_unexpected_fields_stays_fail_closed(self) -> None:
result, _stderr, _request = self.run_client(
reply={"ok": False, "code": "OBSERVATION_UNAVAILABLE", "unexpected": True}
)
self.assertEqual(result, 2)
def test_non_boolean_ok_values_stay_fail_closed(self) -> None:
for reply in (
{"ok": 1},
{"ok": 0, "code": "OBSERVATION_UNAVAILABLE"},
):
with self.subTest(reply=reply):
result, _stderr, _request = self.run_client(reply=reply)
self.assertEqual(result, 2)
if __name__ == "__main__":
unittest.main()
@@ -196,7 +196,12 @@ raise SystemExit(0 if len(session_id) == 64 and denied else 1)
env: { env: {
...process.env, ...process.env,
PATH: `${binDir}:${process.env.PATH ?? ''}`, PATH: `${binDir}:${process.env.PATH ?? ''}`,
MOSAIC_HOME: frameworkRoot, // Point MOSAIC_HOME at the per-entry temp root, NOT the shipped framework
// tree: launch-runtime.py appends its launch ledger to
// $MOSAIC_HOME/fleet/run/sessions/events.ndjson, and writing that into
// framework/ pollutes the tree manifest.spec.ts walks. Nothing in the
// launch chain resolves tools via MOSAIC_HOME (scripts use SCRIPT_DIR).
MOSAIC_HOME: root,
MOSAIC_PRDY_RUNTIME: 'claude', MOSAIC_PRDY_RUNTIME: 'claude',
MOSAIC_LEASE_BROKER_SOCKET: socket, MOSAIC_LEASE_BROKER_SOCKET: socket,
MOSAIC_RUNTIME_GENERATION: '1', MOSAIC_RUNTIME_GENERATION: '1',
@@ -388,8 +393,12 @@ describe('whole mutator-class lease gate', () => {
['claude', 'Write'], ['claude', 'Write'],
['claude', 'NotebookEdit'], ['claude', 'NotebookEdit'],
['claude', 'mcp__provider__close_issue'], ['claude', 'mcp__provider__close_issue'],
['claude', 'Ls'],
['claude', 'Find'],
['pi', 'bash'], ['pi', 'bash'],
['pi', 'edit'], ['pi', 'edit'],
['pi', 'grep'],
['pi', 'find'],
['pi', 'write'], ['pi', 'write'],
['pi', 'deploy'], ['pi', 'deploy'],
['pi', 'unknown_custom_tool'], ['pi', 'unknown_custom_tool'],
@@ -409,8 +418,9 @@ describe('whole mutator-class lease gate', () => {
for (const [runtime, toolName] of [ for (const [runtime, toolName] of [
['claude', 'Read'], ['claude', 'Read'],
['claude', 'Grep'], ['claude', 'Grep'],
['claude', 'Glob'],
['pi', 'read'], ['pi', 'read'],
['pi', 'grep'], ['pi', 'ls'],
['pi', 'mosaic_context_recover'], ['pi', 'mosaic_context_recover'],
] as const) { ] as const) {
expect(await authorize(socket, sessionId, runtime, toolName)).toMatchObject({ expect(await authorize(socket, sessionId, runtime, toolName)).toMatchObject({
+6 -10
View File
@@ -68,8 +68,6 @@ export async function quickStartPath(
// Gateway config + bootstrap // Gateway config + bootstrap
if (!options.skipGateway) { if (!options.skipGateway) {
const headlessRun = process.env['MOSAIC_ASSUME_YES'] === '1' || !process.stdin.isTTY;
try { try {
const configResult = await gatewayConfigStage(prompter, state, { const configResult = await gatewayConfigStage(prompter, state, {
host: options.gatewayHost ?? 'localhost', host: options.gatewayHost ?? 'localhost',
@@ -81,11 +79,9 @@ export async function quickStartPath(
}); });
if (!configResult.ready || !configResult.host || !configResult.port) { if (!configResult.ready || !configResult.host || !configResult.port) {
if (headlessRun) { const message = 'Gateway configuration failed — aborting wizard.';
prompter.warn('Gateway configuration failed in headless mode — aborting wizard.'); prompter.warn(message);
process.exit(1); throw new Error(message);
}
return;
} }
const bootstrapResult = await gatewayBootstrapStage(prompter, state, { const bootstrapResult = await gatewayBootstrapStage(prompter, state, {
@@ -93,9 +89,9 @@ export async function quickStartPath(
port: configResult.port, port: configResult.port,
}); });
if (!bootstrapResult.completed) { if (!bootstrapResult.completed) {
prompter.warn('Admin bootstrap failed — aborting wizard.'); const message = 'Admin bootstrap failed — aborting wizard.';
process.exit(1); prompter.warn(message);
return; throw new Error(message);
} }
finalizeResult.showSummary(); finalizeResult.showSummary();
} catch (err) { } catch (err) {
+22 -16
View File
@@ -348,18 +348,21 @@ async function runFinishPath(
providerType: state.providerType, providerType: state.providerType,
}); });
if (configResult.ready && configResult.host && configResult.port) { if (!configResult.ready || !configResult.host || !configResult.port) {
const message = 'Gateway configuration failed — aborting wizard.';
prompter.warn(message);
throw new Error(message);
}
const bootstrapResult = await gatewayBootstrapStage(prompter, state, { const bootstrapResult = await gatewayBootstrapStage(prompter, state, {
host: configResult.host, host: configResult.host,
port: configResult.port, port: configResult.port,
}); });
if (!bootstrapResult.completed) { if (!bootstrapResult.completed) {
prompter.warn('Admin bootstrap failed — aborting wizard.'); const message = 'Admin bootstrap failed — aborting wizard.';
process.exit(1); prompter.warn(message);
return; throw new Error(message);
} }
finalizeResult.showSummary(); finalizeResult.showSummary();
}
} catch (err) { } catch (err) {
prompter.warn(`Gateway setup failed: ${err instanceof Error ? err.message : String(err)}`); prompter.warn(`Gateway setup failed: ${err instanceof Error ? err.message : String(err)}`);
throw err; throw err;
@@ -420,9 +423,9 @@ async function runHeadlessPath(
}); });
if (!configResult.ready || !configResult.host || !configResult.port) { if (!configResult.ready || !configResult.host || !configResult.port) {
prompter.warn('Gateway configuration failed in headless mode — aborting wizard.'); const message = 'Gateway configuration failed in headless mode — aborting wizard.';
process.exit(1); prompter.warn(message);
return; throw new Error(message);
} }
const bootstrapResult = await gatewayBootstrapStage(prompter, state, { const bootstrapResult = await gatewayBootstrapStage(prompter, state, {
@@ -430,9 +433,9 @@ async function runHeadlessPath(
port: configResult.port, port: configResult.port,
}); });
if (!bootstrapResult.completed) { if (!bootstrapResult.completed) {
prompter.warn('Admin bootstrap failed — aborting wizard.'); const message = 'Admin bootstrap failed — aborting wizard.';
process.exit(1); prompter.warn(message);
return; throw new Error(message);
} }
finalizeResult.showSummary(); finalizeResult.showSummary();
} catch (err) { } catch (err) {
@@ -477,18 +480,21 @@ async function runKeepPath(
skipInstall: options.skipGatewayNpmInstall, skipInstall: options.skipGatewayNpmInstall,
}); });
if (configResult.ready && configResult.host && configResult.port) { if (!configResult.ready || !configResult.host || !configResult.port) {
const message = 'Gateway configuration failed — aborting wizard.';
prompter.warn(message);
throw new Error(message);
}
const bootstrapResult = await gatewayBootstrapStage(prompter, state, { const bootstrapResult = await gatewayBootstrapStage(prompter, state, {
host: configResult.host, host: configResult.host,
port: configResult.port, port: configResult.port,
}); });
if (!bootstrapResult.completed) { if (!bootstrapResult.completed) {
prompter.warn('Admin bootstrap failed — aborting wizard.'); const message = 'Admin bootstrap failed — aborting wizard.';
process.exit(1); prompter.warn(message);
return; throw new Error(message);
} }
finalizeResult.showSummary(); finalizeResult.showSummary();
}
} catch (err) { } catch (err) {
prompter.warn(`Gateway setup failed: ${err instanceof Error ? err.message : String(err)}`); prompter.warn(`Gateway setup failed: ${err instanceof Error ? err.message : String(err)}`);
throw err; throw err;
+3
View File
@@ -253,6 +253,9 @@ importers:
'@mosaicstack/design-tokens': '@mosaicstack/design-tokens':
specifier: workspace:^ specifier: workspace:^
version: link:../../packages/design-tokens version: link:../../packages/design-tokens
'@mosaicstack/types':
specifier: workspace:^
version: link:../../packages/types
better-auth: better-auth:
specifier: ^1.5.5 specifier: ^1.5.5
version: 1.5.5([email protected])([email protected])([email protected](@electric-sql/[email protected])(@opentelemetry/[email protected])(@types/[email protected])(@types/[email protected])([email protected])([email protected])([email protected]))([email protected]([email protected]))([email protected](@opentelemetry/[email protected])(@playwright/[email protected])([email protected]([email protected]))([email protected]))([email protected]([email protected]))([email protected])([email protected](@types/[email protected])(@types/[email protected])([email protected](@noble/[email protected]))([email protected])) version: 1.5.5([email protected])([email protected])([email protected](@electric-sql/[email protected])(@opentelemetry/[email protected])(@types/[email protected])(@types/[email protected])([email protected])([email protected])([email protected]))([email protected]([email protected]))([email protected](@opentelemetry/[email protected])(@playwright/[email protected])([email protected]([email protected]))([email protected]))([email protected]([email protected]))([email protected])([email protected](@types/[email protected])(@types/[email protected])([email protected](@noble/[email protected]))([email protected]))
+31
View File
@@ -12,6 +12,23 @@ MOSAIC_HOME="$TMP/mosaic"
STATE="$TMP/state" STATE="$TMP/state"
LOG="$TMP/npm.log" LOG="$TMP/npm.log"
mkdir -p "$FAKE_BIN" "$HOME_DIR" "$STATE" mkdir -p "$FAKE_BIN" "$HOME_DIR" "$STATE"
REAL_NODE="$(command -v node)"
export MOSAIC_TEST_REAL_NODE="$REAL_NODE"
cat > "$FAKE_BIN/node" <<'FAKE_NODE'
#!/usr/bin/env bash
set -euo pipefail
if [[ "$*" == *'process.versions.node.split'* ]]; then
printf '%s' "${MOSAIC_TEST_NODE_MAJOR:-22}"
exit 0
fi
if [[ "${1:-}" == "--version" ]]; then
printf 'v%s.0.0\n' "${MOSAIC_TEST_NODE_MAJOR:-22}"
exit 0
fi
exec "${MOSAIC_TEST_REAL_NODE:?}" "$@"
FAKE_NODE
chmod +x "$FAKE_BIN/node"
cat > "$FAKE_BIN/npm" <<'FAKE_NPM' cat > "$FAKE_BIN/npm" <<'FAKE_NPM'
#!/usr/bin/env bash #!/usr/bin/env bash
@@ -135,6 +152,20 @@ reset_state() {
rm -f "$STATE"/* rm -f "$STATE"/*
} }
reset_state
echo "[test] --next rejects Node 20 before any install action"
if OUTPUT="$(
HOME="$HOME_DIR" MOSAIC_HOME="$MOSAIC_HOME" MOSAIC_PREFIX="$PREFIX" MOSAIC_NO_COLOR=1 \
MOSAIC_TEST_NPM_LOG="$LOG" MOSAIC_TEST_STATE="$STATE" MOSAIC_TEST_REAL_NODE="$REAL_NODE" \
MOSAIC_TEST_NODE_MAJOR=20 PATH="$FAKE_BIN:$PATH" \
bash "$ROOT/tools/install.sh" --cli --next --yes --no-auto-launch 2>&1
)"; then
echo "expected Node 20 next-lane install to fail" >&2
exit 1
fi
grep -qF 'Node.js >= 22 required for the --next lane' <<<"$OUTPUT"
[[ ! -s "$LOG" ]] || { echo "Node 20 gate ran npm actions" >&2; exit 1; }
reset_state reset_state
echo "[test] --next fast path pins resolved package versions" echo "[test] --next fast path pins resolved package versions"
OUTPUT="$( OUTPUT="$(
+8 -1
View File
@@ -526,6 +526,10 @@ if [[ "$NODE_MAJOR" -lt 20 ]]; then
fail "Node.js >= 20 required (found v$(node --version))" fail "Node.js >= 20 required (found v$(node --version))"
exit 1 exit 1
fi fi
if [[ "$FLAG_NEXT" == "true" && "$NODE_MAJOR" -lt 22 ]]; then
fail "Node.js >= 22 required for the --next lane (found v$(node --version))"
exit 1
fi
echo "" echo ""
echo "${BOLD}Mosaic Stack Installer${RESET}" echo "${BOLD}Mosaic Stack Installer${RESET}"
@@ -762,9 +766,12 @@ if [[ "$FLAG_CHECK" == "false" ]]; then
if "$MOSAIC_CMD" wizard; then if "$MOSAIC_CMD" wizard; then
ok "Wizard complete." ok "Wizard complete."
else else
warn "Wizard exited non-zero." fail "Wizard failed; installation is incomplete."
echo " Completed: framework and CLI installation"
echo " Failed: gateway configuration or admin bootstrap"
echo " You can retry with: ${C}mosaic wizard${RESET}" echo " You can retry with: ${C}mosaic wizard${RESET}"
echo " Or run gateway install alone: ${C}mosaic gateway install${RESET}" echo " Or run gateway install alone: ${C}mosaic gateway install${RESET}"
exit 1
fi fi
fi fi
else else