npm notice run test
npm notice run node --test tests/*.test.mjs
✔ sessionModel: agent vars win, then the system's execution settings (16.0406ms)
✔ a pi bundle: prompt, policy, tools and manifest, 0600 in a 0700 directory (5.910436ms)
✔ a chief-of-staff bundle lists launch among the bus tools and the manifest's typed tools (3.722286ms)
✔ a claude-code bundle adds the wrapped gate hook and the MCP config (3.718002ms)
✔ the prompt's built-in tools line is the harness's own names, as the manifest lists them (14.177234ms)
✔ the prompt names the project clones in the workspace, and none when there are none (8.613524ms)
✔ the prompt explains the review loop to a role with push_working, and reviews to one with fetch_review (10.56918ms)
✔ a bundle is written once: an existing file refuses (2.894341ms)
✔ a path with a single quote can't go into the hook command (5.905444ms)
✔ allow exits 0, a deny exits 2 with the reason on stderr (171.045806ms)
✔ a missing or wrong policy, or a bad event, exits 2 (108.264412ms)
✔ the bundle's wrapped command: a missing gate or node still blocks (1114.264501ms)
✔ claude: typed tools through MCP, the hook blocks, builtins outside --tools don't exist (885.325074ms)
✔ claude: the hook alone blocks a path outside the workspace (451.035657ms)
✔ claude: a second turn resumes the first turn's session (762.911954ms)
✔ claude adapter: --restricted is always passed (5.766707ms)
✔ claude adapter: MOSAIC_CLAUDE_BIN wins over PATH and must be an absolute executable (18.624009ms)
✔ claude adapter: a directory pin exits 2 before it runs; a pin with a space runs (6.213003ms)
✔ claude: CLAUDE.md files and auto-memory don't reach the model; without --restricted they do (755.529814ms)
✔ claude: a missing hook or MCP file refuses before claude starts (7.660741ms)
✔ pi: policy tools and typed tools pass, anything else is blocked (3.804236ms)
✔ claude: builtins map from pi names, typed tools need the mcp prefix (0.833357ms)
✔ file tool paths must resolve inside the workspace (2.139432ms)
✔ pi's own path normalisation can't be used to step out (1.465516ms)
✔ a symlink inside the workspace that points out is outside (1.213249ms)
✔ a dangling symlink is refused at any depth, in both harnesses (5.031188ms)
✔ read is checked under every spelling pi's read would open, in both harnesses (19.801462ms)
✔ other spellings cover directories, dangling links and pi's cwd (2.431906ms)
✔ a relative path climbs from the workspace's real path, in both harnesses (7.789826ms)
✔ claude path fields per tool (1.277648ms)
✔ glob patterns stay inside the workspace (1.356561ms)
✔ a path that can't be checked is blocked (1.223306ms)
✔ fetch_review's have lists the working refs before origin's, and no more than HAVE_MAX (324.707844ms)
✔ fetch_review's have orders the working refs newest first within each clone, one clone after another (175.812599ms)
✔ a runner git past its time: SIGTERM reaches its process group, and SIGKILL follows after the grace period (1809.248133ms)
✔ once a signalled runner git has closed, its group gets SIGKILL, so a process that ignored SIGTERM dies too (303.908368ms)
✔ initialize, ping and tools/list (55.590376ms)
✔ tools/call goes through the tool socket; a refusal is an isError result (48.590303ms)
✔ replies are compact JSON; over the cap a tool error names the size and the cap (39.900191ms)
✔ unknown tools and methods are JSON-RPC errors and never reach the socket (47.214535ms)
✔ a missing argument is a usage error (33.257172ms)
✔ pi: typed tools reach the socket, the gate blocks, agent_end writes the marker (459.574012ms)
✔ pi: a reply over the cap is a tool error naming the size and the cap; list_tasks asks for open tasks (357.34943ms)
✔ pi: a write through a dangling symlink is blocked, and nothing appears outside (338.750236ms)
✔ pi: a read is refused when pi would open another spelling outside (351.196051ms)
✔ pi: a relative path climbs from the real path of a workspace behind a symlink (340.036647ms)
✔ pi: a relative path climbs from the real path of a dataRoot behind a symlink (338.839081ms)
✔ pi: a missing extension refuses before any model call (6.953889ms)
✔ pi: an extension without its configuration fails pi's start (265.450423ms)
✔ founderCheck: founder variables, then a needed service without a usable token (1.53339ms)
✔ turnRequest names the sender, class, reply and decision (0.26491ms)
✔ turnRequest says a tester's message is a test and never Jason's instruction (0.307269ms)
✔ turnRequest shows the message's own sent time, a day old as well as new (0.15313ms)
✔ stateBlock names the launch state with its time and event, and the open tasks with the cap (0.356621ms)
✔ a message becomes a turn, the answer goes back as a RESULT, SIGTERM releases and exits 0 (222.829101ms)
✔ launches revoked and then restored: the turn's block reads allowed and gives the restore time (198.560674ms)
✔ a task closed on the tracker is missing from the next turn's open list (196.280923ms)
✔ a message read a day after it was sent shows the time it was sent (114.35617ms)
✔ a decision the human resolves becomes a turn citing the decision and the answer (141.006849ms)
✔ a failed state read fails the turn: no adapter runs, the sender hears why, the message is read (63.506904ms)
✔ a SIGTERM before the claim stops the runner with exit 0 and no claim (50.569349ms)
✔ typed tools carry the runner's capability; launch goes to the host's launch socket (298.048233ms)
✔ a RESULT gets no automatic reply; failed turns reply with the reason (1215.668498ms)
✔ SIGTERM during a turn kills the turn's process group and still exits 0 (93.954721ms)
✔ founder credentials stop before the claim (20) (95.422471ms)
✔ a refused claim exits 21; an ended run's capability exits 22 (103.02915ms)
✔ the launch ending under a running session exits 22 (99.187605ms)
✔ a broker that stays unreachable exits 23 after brokerRetries polls (202.148293ms)
✔ a broker that is down at the claim exits 23, not 21 (58.844579ms)
✔ no capability, or a malformed one, on stdin exits 2 (158.388419ms)
✔ a missing or malformed policy exits 2 before the claim (80.058072ms)
✔ the PM gets its task verbs and the reads, and no launch (8.896094ms)
✔ the chief of staff gets launch and send_message, and no task verbs (2.87129ms)
✔ the assistant gets send_message and the reads, no launch (3.97893ms)
✔ list_requirements reads the requirement ids, for a task.create holder with a tracker (5.113241ms)
✔ a coder gets no launch, no resolve_decision, and no task tools without a tracker (1.961387ms)
✔ a coder with a tracker gets its own task tools and no task_reassign (1.2986ms)
✔ launch only when the business's launch block names the instance as launcher (2.788318ms)
✔ an action outside the instance's authority has no tool (2.58988ms)
✔ the review loop: push_working and request_review for the coder, fetch_review and review_verdict for the reviewer, none for the PM (9.17938ms)
✔ callTool: one JSON line out, the result back, a refusal rejects (18.568899ms)
✔ list_tasks takes the broker's filters and pages; get_task takes a task_ref (3.405153ms)
✔ callTool sends list_tasks done false unless the session names done; other tools as given (4.070951ms)
✔ replyText: compact JSON up to the cap, an error naming the size and the cap over it (1.078556ms)
ℹ tests 84
ℹ suites 0
ℹ pass 84
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 10739.604008
