✔ sessionModel: agent vars win, then the system's execution settings (13.041523ms)
✔ a pi bundle: prompt, policy, tools and manifest, 0600 in a 0700 directory (4.403863ms)
✔ a chief-of-staff bundle lists launch among the bus tools and the manifest's typed tools (3.405445ms)
✔ a chief-of-staff bundle lists end_run and change_organization as typed and nothing untyped, for either harness (6.470466ms)
✔ the stop line names end_run for a role.end holder, and every other role keeps today's text (7.34616ms)
✔ a claude-code bundle adds the wrapped gate hook and the MCP config (2.187121ms)
✔ the prompt's built-in tools line is the harness's own names, as the manifest lists them (7.627295ms)
✔ the prompt names the project clones in the workspace, and none when there are none (6.224544ms)
✔ the prompt explains the review loop to a role with push_working, and reviews to one with fetch_review (6.323724ms)
✔ a bundle is written once: an existing file refuses (1.704039ms)
✔ a path with a single quote can't go into the hook command (1.795961ms)
✔ allow exits 0, a deny exits 2 with the reason on stderr (123.80605ms)
✔ a missing or wrong policy, or a bad event, exits 2 (82.141073ms)
✔ the bundle's wrapped command: a missing gate or node still blocks (1088.303104ms)
✔ claude: typed tools through MCP, the hook blocks, builtins outside --tools don't exist (724.192724ms)
✔ claude: the hook alone blocks a path outside the workspace (397.544697ms)
✔ claude: a second turn resumes the first turn's session (687.37367ms)
✔ claude adapter: --restricted is always passed (5.268451ms)
✔ claude adapter: MOSAIC_CLAUDE_BIN wins over PATH and must be an absolute executable (22.876557ms)
✔ claude adapter: a directory pin exits 2 before it runs; a pin with a space runs (5.672402ms)
✔ claude: CLAUDE.md files and auto-memory don't reach the model; without --restricted they do (675.74951ms)
✔ claude: a missing hook or MCP file refuses before claude starts (7.560051ms)
✔ pi: policy tools and typed tools pass, anything else is blocked (3.634648ms)
✔ claude: builtins map from pi names, typed tools need the mcp prefix (0.700764ms)
✔ file tool paths must resolve inside the workspace (1.257459ms)
✔ pi's own path normalisation can't be used to step out (0.987186ms)
✔ a symlink inside the workspace that points out is outside (0.744498ms)
✔ a dangling symlink is refused at any depth, in both harnesses (2.502264ms)
✔ read is checked under every spelling pi's read would open, in both harnesses (16.084446ms)
✔ other spellings cover directories, dangling links and pi's cwd (2.229128ms)
✔ a relative path climbs from the workspace's real path, in both harnesses (7.205336ms)
✔ claude path fields per tool (0.858695ms)
✔ glob patterns stay inside the workspace (0.818252ms)
✔ a path that can't be checked is blocked (0.501085ms)
✔ fetch_review's have lists the working refs before origin's, and no more than HAVE_MAX (205.251265ms)
✔ fetch_review's have orders the working refs newest first within each clone, one clone after another (132.727971ms)
✔ a runner git past its time: SIGTERM reaches its process group, and SIGKILL follows after the grace period (1808.736695ms)
✔ once a signalled runner git has closed, its group gets SIGKILL, so a process that ignored SIGTERM dies too (304.203842ms)
✔ initialize, ping and tools/list (54.278981ms)
✔ tools/call goes through the tool socket; a refusal is an isError result (36.295281ms)
✔ replies are compact JSON; over the cap a tool error names the size and the cap (34.422659ms)
✔ unknown tools and methods are JSON-RPC errors and never reach the socket (28.715132ms)
✔ a missing argument is a usage error (28.229512ms)
✔ pi: typed tools reach the socket, the gate blocks, agent_end writes the marker (357.451816ms)
✔ pi: a reply over the cap is a tool error naming the size and the cap; list_tasks asks for open tasks (333.525668ms)
✔ pi: a write through a dangling symlink is blocked, and nothing appears outside (306.126441ms)
✔ pi: a read is refused when pi would open another spelling outside (312.158201ms)
✔ pi: a relative path climbs from the real path of a workspace behind a symlink (313.988787ms)
✔ pi: a relative path climbs from the real path of a dataRoot behind a symlink (300.31232ms)
✔ pi: a missing extension refuses before any model call (5.948324ms)
✔ pi: an extension without its configuration fails pi's start (246.889577ms)
✔ founderCheck: founder variables, then a needed service without a usable token (1.684609ms)
✔ turnRequest names the sender, class, reply and decision (0.298978ms)
✔ turnRequest says a tester's message is a test and never Jason's instruction (0.302168ms)
✔ turnRequest shows the message's own sent time, a day old as well as new (0.192261ms)
✔ stateBlock names the launch state with its time and event, and the open tasks with the cap (0.334532ms)
✔ a message becomes a turn, the answer goes back as a RESULT, SIGTERM releases and exits 0 (211.830002ms)
✔ launches revoked and then restored: the turn's block reads allowed and gives the restore time (168.599655ms)
✔ a task closed on the tracker is missing from the next turn's open list (168.920157ms)
✔ a message read a day after it was sent shows the time it was sent (115.722138ms)
✔ a decision the human resolves becomes a turn citing the decision and the answer (139.291876ms)
✔ a failed state read fails the turn: no adapter runs, the sender hears why, the message is read (62.561616ms)
✔ a SIGTERM before the claim stops the runner with exit 0 and no claim (48.35055ms)
✔ typed tools carry the runner's capability; launch goes to the host's launch socket (273.930492ms)
✔ end_run and change_organization reach the broker with the tool's arguments, and refusals keep their codes (278.596797ms)
✔ a RESULT gets no automatic reply; failed turns reply with the reason (1186.978066ms)
✔ SIGTERM during a turn kills the turn's process group and still exits 0 (93.401135ms)
✔ founder credentials stop before the claim (20) (86.705147ms)
✔ a refused claim exits 21; an ended run's capability exits 22 (88.963519ms)
✔ the launch ending under a running session exits 22 (98.846955ms)
✔ a broker that stays unreachable exits 23 after brokerRetries polls (198.750251ms)
✔ a broker that is down at the claim exits 23, not 21 (46.126166ms)
✔ no capability, or a malformed one, on stdin exits 2 (155.111303ms)
✔ a missing or malformed policy exits 2 before the claim (88.028014ms)
✔ the PM gets its task verbs and the reads, and no launch (7.951473ms)
✔ the chief of staff gets launch, send_message, end_run and change_organization, and no task verbs (2.535761ms)
✔ end_run sends role and reason, no decision; change_organization sends revision and the organization keys (2.315877ms)
✔ the organization keys change_organization takes are the broker's ORGANIZATION_KEYS (0.11551ms)
✔ a role without role.end or organization.change gets neither tool, and each tool follows its own grant (9.289619ms)
✔ the assistant gets send_message and the reads, no launch (2.281063ms)
✔ list_requirements reads the requirement ids, for a task.create holder with a tracker (2.46255ms)
✔ a coder gets no launch, no resolve_decision, and no task tools without a tracker (0.972783ms)
✔ a coder with a tracker gets its own task tools and no task_reassign (0.88871ms)
✔ launch only when the business's launch block names the instance as launcher (1.855726ms)
✔ an action outside the instance's authority has no tool (1.240953ms)
✔ the review loop: push_working and request_review for the coder, fetch_review and review_verdict for the reviewer, none for the PM (3.786528ms)
✔ callTool: one JSON line out, the result back, a refusal rejects (10.332668ms)
✔ list_tasks takes the broker's filters and pages; get_task takes a task_ref (3.05533ms)
✔ callTool sends list_tasks done false unless the session names done; other tools as given (3.029844ms)
✔ replyText: compact JSON up to the cap, an error naming the size and the cap over it (0.588526ms)
ℹ tests 90
ℹ suites 0
ℹ pass 90
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 10883.303765
