✔ W1: two processes acquire the same pair at once; exactly one claim (209.812305ms)
✔ W1: two writers publish the same revision at once: one wins, the other gets null, the winner's record stays (31.596713ms)
✔ W1: a revision name appears only after its bytes are synced; before that, only a temp file exists (16.865413ms)
✔ W2: acquire while a claim is reserved or active refuses already-active (409.009847ms)
✔ W3: acquire while stopping, uncertain, or stopped without proof refuses unsafe-replacement (382.734026ms)
✔ W4: same session with another seat tuple, and the reverse, both refuse; a loser on the seat key closes it no-unit (234.151608ms)
✔ W4: a hard link of one session under another seat is the same session: the second controller refuses already-active and launches nothing (198.261514ms)
✔ W4: a copy of one session under another seat is the same session: the second controller refuses already-active and launches nothing (190.708869ms)
✔ W4: a session header ID that changes after construction refuses target; nothing is claimed or launched (2.948471ms)
✔ W5: SIGKILL between every publication barrier of acquire and transition; restart never finds two holders or a lost claim (11976.054483ms)
✔ W5: SIGKILL between every publication barrier of release; restart finishes or holds the release (39366.094305ms)
✔ W6: controller killed mid-turn while the engine lives; restart is uncertain, no launch, prompts refuse (262.918616ms)
✔ W12: a live owner paused with SIGSTOP; a second controller refuses already-active and changes nothing (143.497175ms)
✔ W13: crash after the engine spawns, before active; restart finds the live unit: uncertain, no second spawn, force stop only (616.11398ms)
✔ W14: crash after reservation, before the spawn marker: stopped with a no-unit observation; the pair is free (279.842443ms)
✔ W20: crash after the spawn marker, scope collected; uncertain in both runs, the marker is copied, no launch until a boot proof (359.418164ms)
✔ W15: crash between the two keys during release; restart finishes it under the same claim ID (112.281746ms)
✔ W7: recorded boot ID differs on the same machine: stopped with a boot proof; open tool calls become uncertain (215.496192ms)
✔ W8: resume after a proven stop with the same pins: new claim ID, generation +1, same conversation, branch and leaf (270.262984ms)
✔ W9: resume with a changed binary, argv digest, branch or leaf is refused and the claim is unchanged (754.103211ms)
✔ W11: the controller writes no session file; only the fake engine's own appends appear (259.694804ms)
✔ W16: a highest revision that won't parse holds the pair uncertain; the older stopped revision is not reused (91.528696ms)
✔ W17: a claim root copied from another host refuses foreign-host and promotes nothing (72.767099ms)
✔ G1: a session path or claim root under .pi/state, ~/.claude, the data root or a registration refuses at construction (3.52146ms)
✔ G2: a symlink inside the fixture root to a live session file is refused by the real-path check (1.098365ms)
✔ G3: a fixture path swapped for a live path after construction is refused at bind (1.791037ms)
✔ no shim from this file's tests is left running for the after hook (29.227788ms)
✔ K1: force stop kills a tool child that called setsid; stopped with a verified proof (1035.074876ms)
✔ K2: K1 on the process-group fallback ends uncertain, never stopped (444.063051ms)
✔ K3: SIGTERM acknowledged while a member lives: stopping until the kill phase, never stopped from TERM (747.970793ms)
✔ K4: two engines; force stop one; the other survives by independent observation (2975.7536ms)
✔ K5: a stop during a tool call leaves the effect uncertain, and it is shown (679.355826ms)
✔ K12: a member forking in a loop: the freeze stops it, enumeration is complete, populated 0 after cgroup.kill (830.856437ms)
✔ K13: a member writing its pid into another cgroup is refused by the namespace; the kill is complete (833.005125ms)
✔ K15: the shim gone, engine/cgroup.events unreadable, or the engine cgroup missing: evidence unavailable, not empty; uncertain (5756.682073ms)
✔ K10: controller killed between the TERM and kill phases: restart checks the invocation ID and re-runs from TERM for the same stop (768.039865ms)
✔ K11: controller killed after the confirmation is recorded, before TERM: restart checks the invocation ID and re-runs from TERM for the same stop (585.80089ms)
✔ K14: a unit with the recorded name but another invocation ID: evidence unavailable, no signals, uncertain (659.24289ms)
✔ K6: recover without proof, without confirmation, or with changed pins is refused (613.572007ms)
✔ K7: recover after proof, then launch: new claim and execution, generation +1, same leaf; the cancelled prompt is not replayed (315.870588ms)
✔ K8: an engine that loads another leaf on resume is refused before admission; it stays claimed until a proven stop (452.34197ms)
✔ K9: an interrupt that never settles stays uncertain; force stop stays available; takeover is refused while fenced (3285.257435ms)
✔ K16: a claim from another machine ID refuses foreign-host; no boot proof is issued (120.380693ms)
✔ K17: two launcher calls with one eligibility record: one launch, the other refuses, no second engine (381.635805ms)
✔ K18: the leaf changes after eligibility: launch refused; the reservation stays until released with proof (375.473852ms)
✔ K19: a scope launched with only the engine environment still reaches the user manager; the engine sees no other names (86.767929ms)
✔ R1: after a proven force stop the controller releases the scope; the shim and the unit are gone (478.110743ms)
✔ R2: close releases the scope of a binding proven stopped, once, when the stop's own release hasn't run (656.340734ms)
✔ R3: no release after an unavailable stop or an unverified proof: the scope and its shim stay as evidence (1585.899954ms)
✔ R4: after a normal engine exit the scope stays until a proven force stop releases it (847.578589ms)
✔ R5: the harness finds a running shim and reaps it, and never signals a shim outside a mosaic-chat- scope (Filbert N1 and N2 on #1533) (448.219879ms)
✔ R6: releaseCohort releases nothing while the engine runs, for another invocation ID, or for a cohort without a scope (318.640551ms)
✔ R7: controller killed after the claim is stopped, before the release: the restart releases the scope (3026.507972ms)
✔ R8: controller killed between the two keys' stopped revisions: the restart releases the scope (552.965142ms)
✔ R9: a release that ended unavailable is tried again on a confirmed recover, not a refused one (704.598562ms)
✔ R10: close releases and signals nothing for a stopped binding whose proof no longer verifies (Filbert N3 on #1536) (657.20104ms)
✖ R11: close({ killEngine: true }) after a proven stop sends no signal to the recorded engine PID; a restart releases nothing more (2541.711065ms)
✔ R12: a restart onto a claim stopped on a boot proof releases nothing, though its scope is listed (273.218951ms)
✔ no shim from this file's tests is left running for the after hook (24.675367ms)
✔ S1: `/goal x`, with leading spaces or a tab, refuses text-policy at admission; zero engine bytes (203.278494ms)
✔ S2: every prefix pinned Pi interprets is refused, from the list the code uses; the rest reach the engine exactly (190.159429ms)
✔ S3: `/goal` on the second line is pinned from the source: Pi checks only index 0, so it is admitted and sent exactly (284.165594ms)
✔ S4: a `/` left in the composer is cleared when control transfers and returns; the next submit sends only the new text (276.216987ms)
✔ S5: an observer terminal gets a paste then Enter, as send-message.sh does: not admitted: controller, nothing sent (192.580108ms)
✔ S6: a mediated-shaped registration (no tmux) passed to the board's replyToRow: 409 no tmux session; exec never runs (0.427424ms)
✔ S7: ESC, bracketed-paste markers and U+2028/U+2029 travel as one JSON string; the engine receives the exact text in one record (198.125481ms)
✔ P3: a Pi confirm, select, input or editor dialog is shown disabled with a reason and never answered (306.1816ms)
✔ E1: send, ack, user, toolCall, toolResult, final answer: shown once, no refresh, draft and reading position kept (205.357906ms)
✔ E2: U+2028, U+2029 inside JSON strings and CRLF line ends each parse as one record, on the splitter and through the controller (171.935943ms)
✔ E3: a multipart final, two blocks, null request correlation and duplicate delivery (203.4889ms)
✔ E4: a page read after message_end but before its entry is persisted: marker at the seam, re-read after run-settled, each message once (190.720463ms)
✔ E4: a gap or a new epoch also reconciles; nothing is concatenated across a gap (11.07304ms)
✔ E5: an unknown native event gives no client event; evidence records its type and bytes; the terminal count goes up (193.427393ms)
✔ E6: a tool result delayed across a pause and a reconnect is reconciled without a manual refresh (213.752328ms)
✔ E7: the terminal renders the same stream as the library client, as observer and then as controller, and submits only as controller (234.238319ms)
✔ terminal: engine control characters are made visible; a lost connection refuses submit (189.47783ms)
✔ terminal: outcome unknown is shown as such, with no resend offer, and nothing is resent (0.396529ms)
✔ terminal: text after Enter in the same input chunk starts the next message; it never joins the one submitted (0.287241ms)
✔ terminal: a paste-start marker split right after its ESC still opens the paste; the Enter inside it never submits (0.317257ms)
✔ terminal: Ctrl-T then Enter in one chunk is judged after the takeover, as if typed one key at a time (Filbert F2, #1507) (0.471462ms)
✔ terminal: input held behind Ctrl-T waits for that takeover while an earlier action is still pending (Darkwing T1 on #1522) (30.840346ms)
✔ terminal: a hold inside held input holds again, and once it is drained later input and Ctrl-C still reach the terminal (Darkwing note 1 on #1522) (0.421114ms)
✔ terminal: an action that throws still releases the input held behind it, in order, then rethrows (5.788254ms)
✔ terminal: after an action throws, later input still runs; input() puts the error in the status line (Filbert N1 on #1522) (0.355126ms)
✔ terminal: input() reports an error when it happens, so it never overwrites a later status; held input's promise doesn't carry the holder's error (Filbert N4 on #1522) (0.30335ms)
✔ terminal: invisible and bidi characters are made visible; head, status and notice lines stay one line (0.100639ms)
✔ every record these fixtures produced is a valid CHAT-01 record (E5: no record fails the schema) (217.928978ms)
✔ H1: two takeovers with the same expected generation: one wins, +1; the other refuses generation (244.055458ms)
✔ H2: the old controller's prompt after a takeover commits is refused with zero engine bytes (362.856651ms)
✔ H3: a takeover while a prompt holds the dispatch lock: written under the old actor, or refused; never both (800.119886ms)
✔ H4: self-takeover is refused (196.524786ms)
✔ H9: Interrupt racing a prompt's dispatch: before the write, dispatch-refused and no-turn; after, §3 rules (411.032099ms)
✔ H10: Interrupt and force stop together: one stop chain, force stop supersedes (975.505451ms)
✔ H10: an overlap during the pause before the abort: no abort, the stop ends uncertain (212.751537ms)
✔ H10: a no-turn Interrupt lifts only its own fence; admission stays closed under force stop, overlap or revocation (522.6803ms)
✔ H11: the controller disconnects mid-turn: work continues, the claim is unchanged, control stays put (259.129737ms)
✔ H12: an exact retry after reconnecting to the same incarnation returns the same receipt; one dispatch (120.968325ms)
✔ H13: a retry with the same request ID and different text is refused (167.479335ms)
✔ H14: late stdout from the old engine after a replacement is dropped by incarnation, counted, never rendered (602.934601ms)
✔ H15: a revoked connection's command is refused; the revocation fence holds (272.219954ms)
✔ H16: a second controller for the same session refuses already-active; the first is untouched (200.97538ms)
✔ H10: a second force stop while the first escalation runs refuses fenced; one escalation, and the claim records only the first stop's phases (846.808199ms)
✔ a force stop whose fence throws leaves no escalation flag behind, so the next force stop runs (Darkwing F2, #1507) (445.990408ms)
✔ H17: a confirmation reused, answered from another connection, or used after the stop changed is refused (908.374606ms)
✔ H18: two prompts before any native output: the second refuses busy; one engine write (197.114166ms)
✔ H19: the pipe fails mid-line under a large prompt: delivery-unknown transport-unknown, poisoned, no later write (337.362618ms)
✔ H19: the link itself never writes again after an unknown outcome, whoever calls it (0.675388ms)
✔ H19: the controller dies mid-write of a large line: after restart the outcome is unknown and nothing is resent (661.274751ms)
✔ H20: the line is written but the ack is lost when the controller dies: orphan, outcome unknown, nothing resent (567.551798ms)
✔ H21: a retry of the exact request with the old token after a crash is stale-incarnation; no second write (684.46252ms)
✔ H22: after H21 and a valid recovery, a new request with the new token is admitted (3069.254786ms)
✔ H23: requests pending at a restart are not resent; each shows outcome unknown (591.446983ms)
✔ no shim from this file's tests is left running for the after hook (25.600964ms)
✔ a plain conversation: catalogue row, one page, CHAT-01 records (6.094545ms)
✔ native entries map to blocks: tools, thinking, bash, notices, ids that do not fit (1.67267ms)
✔ F1: a malformed line is an unavailable part at its position, and reading continues (1.800012ms)
✔ F1: a missing parent stops the history with a notice that names the unreadable lines (2.399652ms)
✔ F1: an unreadable fork is never merged into another branch's history (1.511162ms)
✔ F1: a follow stays on its branch when the next entry's parent is unreadable (1.792705ms)
✔ F1: a file whose entries are all unreadable shows a notice per line (0.820786ms)
✔ F2: a truncated trailing line marks the view incomplete, not an error (1.175867ms)
✔ pagination: 100 parts, then the rest; parts concatenate to the whole branch (3.065015ms)
✔ F3: a replaced file (new inode) refuses old cursors with reconcile (4.002963ms)
✔ F4: a same-inode rewrite of the prefix refuses old cursors with reconcile (4.205844ms)
✔ F5: growth between pages keeps the epoch and the page stops at the pinned length (4.321985ms)
✔ F6: unknown, foreign and expired cursors refuse and leave the cursor usable (7.553294ms)
✔ F7: a symlinked file and a symlinked directory component are refused, never opened (8.248563ms)
✔ F8: a file swapped for a symlink after the catalogue is refused (1.899766ms)
✔ F9: registrations never add or redirect a root (1.936515ms)
✔ F10: a header cwd naming another project is refused (4.081673ms)
✔ F11: parentSession renders with a marker and the parent is never opened (0.783219ms)
✔ F12: two leaves: the default leaf is shown and the other branch reads alone (4.2913ms)
✔ F12: a follow refuses when an appended duplicate id changes the branch's earlier parts (2.10679ms)
✔ F12: a second root (Pi's resetLeaf) starts its own branch (1.463133ms)
✔ F13: compaction is a marker in place, then the retained content (0.784714ms)
✔ F14: long strings split into fragments and parts, reassemble exactly, and pages respect the byte cap (533.074989ms)
✔ fragments never cut a surrogate pair and keep an empty string (4.832583ms)
✔ F15: a Claude seat is an unsupported-harness placeholder whose directory is never read (1.651953ms)
✔ unknown conversations, empty files and non-Pi files refuse (2.36422ms)
✔ an unreadable file or root inside the roots is refused per row, not a failed catalogue (1.027547ms)
✔ a seat directory without search permission refuses that root, not the catalogue (1.904937ms)
✔ every page and cursor is a valid CHAT-01 record (755.259232ms)
✔ the engine pin holds for the installed package (1.789637ms)
✔ pinned Pi, sealed and without credentials, answers the controller's commands with the shapes the fake models (276.96992ms)
✔ sealed, pinned Pi ignores a trusted project's .pi resources; --approve past the seal would load them, and checkSeal refuses it (Filbert F2 on #1522) (519.896012ms)
✔ pinned Pi appends thinking_level_change at start when the branch lacks one, so the leaf moves (K8 then fails closed) (253.034583ms)
✔ N25: ordinary Interrupt reconciles; a non-empty queue_update in the window is O5 (542.506214ms)
✔ N1: an extension's follow-up queued after the fence is cleared before any abort; O5, Unknown (441.081989ms)
✔ N1: a follow-up queued before the fence is O5 at once; the Interrupt refuses fenced (229.238773ms)
✔ N2: with abort first, the fake runs the external item (the ordering guard has teeth) (21.793548ms)
✔ N3: the fence lands in preflight, preflight errors, no run: failed, No run, uncertain (351.907095ms)
✔ N4: the ack arrives after the first abort and a run starts: clear and abort again; Interrupted (194.730682ms)
✔ N5: an input handler takes the prompt: ack, no run, delivery-unknown handled-without-run (277.342536ms)
✔ N6: an extension queues between clear_queue and abort: O5 and O6, Unknown (446.459741ms)
✔ N7: clear_queue times out: no abort, nativeQueue unknown, force stop still ends it (1861.831437ms)
✔ N7: clear_queue answers an error: no abort, nativeQueue unknown, the link not poisoned (195.66958ms)
✔ N8: an extension prompt starts a run during Mosaic preflight; the losing settle is O3 (280.564421ms)
✔ N9: a run that started before the fence and ends aborted: failed interrupted, Interrupted (158.344467ms)
✔ N9: decision 34: a run that ends aborted with no stop in progress: aborted-without-stop, uncertain, outcome unknown (174.885146ms)
✔ N9: an aborted that lands after the fence but before any abort is written: aborted-without-stop, Unknown (382.451067ms)
✔ N10: fake conformance (32.692746ms)
✔ N11: the run fails before any user message_start: delivery-unknown ack-without-start, never failed (412.148167ms)
✔ N12: input that starts a run after the final empty clear is O1 and not part of the stop's proof (330.487566ms)
✔ N13: agent_start with no slot held is O1; a later prompt refuses with zero engine bytes (334.684261ms)
✔ N14: the run completes while clear_queue is in flight: finished, Completed first, uncertain (533.167021ms)
✔ N14: the run completes after the abort is written, before Pi applies it: finished, never relabelled (344.461037ms)
✔ N15: the fence lands in preflight, then an input handler takes it: handled-without-run, No run (354.541594ms)
✔ N16: Interrupt with no slot and no run refuses no-turn: no stop, no bytes, admission open (160.860968ms)
✔ N17: the run fails on its own during the exchange: failed, Failed on its own (260.88606ms)
✔ N18: no final assistant message_end, or a lost line: working stays working; before working, transport-unknown (699.367313ms)
✔ N19: a losing extension prompt settles inside the Mosaic run before its user message: O3, run-overlap (483.641113ms)
✔ N20: an extension triggerTurn during Mosaic preflight starts first; while streaming it queues with no signal (444.86105ms)
✔ N21: a losing settle after the receipt settled finished is O2; the receipt stays finished (229.607426ms)
✔ N22: an agent-level custom message is dropped by the clear with no signal; evidence names the seal (174.032052ms)
✔ N23: a nextTurn message survives clear and abort and attaches to the next prompt, with no signal (178.164937ms)
✔ N24: the seal is an allow-list: --extension, a missing --no-* flag, a second --mode or --session, a session or output flag, or a stray word refuses unsealed-engine; no engine starts (188.268429ms)
✔ N24b: the seal covers the engine command and environment: config can't name either, the env is built from names, and a mutated command is refused at bind (130.040069ms)
ℹ tests 177
ℹ suites 0
ℹ pass 176
ℹ fail 1
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 60228.738662

✖ failing tests:

test at packages/conversation/tests/cohort.test.mjs:1081:1
✖ R11: close({ killEngine: true }) after a proven stop sends no signal to the recorded engine PID; a restart releases nothing more (2541.711065ms)
  AssertionError [ERR_ASSERTION]: Expected values to be strictly deep-equal:
  + actual - expected
  
  + [
  +   {
  +     at: '2026-10-10T06:55:17.157Z',
  +     claim: 'c28dd8cd48878438e098ae625',
  +     kind: 'release',
  +     outcome: 'unavailable',
  +     reason: 'the shim socket is gone',
  +     unitName: 'mosaic-chat-c28dd8cd48878438e098ae625',
  +     why: 'start'
  +   }
  + ]
  - []
  
      at TestContext.<anonymous> (file:///home/jwoltje/darkwing-scratch/r51b/wt2/packages/conversation/tests/cohort.test.mjs:1097:12)
      at process.processTicksAndRejections (node:internal/process/task_queues:104:5)
      at async Test.run (node:internal/test_runner/test:1409:7)
      at async Test.processPendingSubtests (node:internal/test_runner/test:974:7) {
    generatedMessage: true,
    code: 'ERR_ASSERTION',
    actual: [ { kind: 'release', claim: 'c28dd8cd48878438e098ae625', unitName: 'mosaic-chat-c28dd8cd48878438e098ae625', why: 'start', outcome: 'unavailable', reason: 'the shim socket is gone', at: '2026-10-10T06:55:17.157Z' } ],
    expected: [],
    operator: 'deepStrictEqual',
    diff: 'simple'
  }
