✔ sessionModel: agent vars win, then the system's execution settings (13.399692ms)
✔ a pi bundle: prompt, policy, tools and manifest, 0600 in a 0700 directory (4.584534ms)
✔ a chief-of-staff bundle lists launch among the bus tools and the manifest's typed tools (2.296631ms)
✔ a claude-code bundle adds the wrapped gate hook and the MCP config (3.053174ms)
✔ the prompt's built-in tools line is the harness's own names, as the manifest lists them (9.249709ms)
✔ the prompt names the project clones in the workspace, and none when there are none (7.272513ms)
✔ the prompt explains the review loop to a role with push_working, and reviews to one with fetch_review (7.638611ms)
✔ a bundle is written once: an existing file refuses (1.460462ms)
✔ a path with a single quote can't go into the hook command (2.389599ms)
✔ allow exits 0, a deny exits 2 with the reason on stderr (125.530748ms)
✔ a missing or wrong policy, or a bad event, exits 2 (74.968745ms)
✔ the bundle's wrapped command: a missing gate or node still blocks (1084.255248ms)
✔ claude: typed tools through MCP, the hook blocks, builtins outside --tools don't exist (706.733004ms)
✔ claude: the hook alone blocks a path outside the workspace (457.849325ms)
✔ claude: a second turn resumes the first turn's session (679.61127ms)
✔ claude adapter: --restricted is always passed (5.13109ms)
✔ claude adapter: MOSAIC_CLAUDE_BIN wins over PATH and must be an absolute executable (16.156421ms)
✔ claude adapter: a directory pin exits 2 before it runs; a pin with a space runs (5.548467ms)
✔ claude: CLAUDE.md files and auto-memory don't reach the model; without --restricted they do (785.931942ms)
✔ claude: a missing hook or MCP file refuses before claude starts (6.918007ms)
✔ pi: policy tools and typed tools pass, anything else is blocked (3.349528ms)
✔ claude: builtins map from pi names, typed tools need the mcp prefix (0.737366ms)
✔ file tool paths must resolve inside the workspace (1.213461ms)
✔ pi's own path normalisation can't be used to step out (1.111981ms)
✔ a symlink inside the workspace that points out is outside (0.969621ms)
✔ a dangling symlink is refused at any depth, in both harnesses (2.411056ms)
✔ read is checked under every spelling pi's read would open, in both harnesses (15.05931ms)
✔ other spellings cover directories, dangling links and pi's cwd (1.761293ms)
✔ a relative path climbs from the workspace's real path, in both harnesses (4.545519ms)
✔ claude path fields per tool (1.123275ms)
✔ glob patterns stay inside the workspace (0.76606ms)
✔ a path that can't be checked is blocked (0.364174ms)
✔ fetch_review's have lists the working refs before origin's, and no more than HAVE_MAX (198.382116ms)
✔ fetch_review's have orders the working refs newest first within each clone, one clone after another (133.129597ms)
✔ a runner git past its time: SIGTERM reaches its process group, and SIGKILL follows after the grace period (1810.286267ms)
✔ once a signalled runner git has closed, its group gets SIGKILL, so a process that ignored SIGTERM dies too (304.639152ms)
✔ initialize, ping and tools/list (47.853177ms)
✔ tools/call goes through the tool socket; a refusal is an isError result (32.118526ms)
✔ replies are compact JSON; over the cap a tool error names the size and the cap (33.501923ms)
✔ unknown tools and methods are JSON-RPC errors and never reach the socket (29.955629ms)
✔ a missing argument is a usage error (26.151906ms)
✔ pi: typed tools reach the socket, the gate blocks, agent_end writes the marker (364.983523ms)
✔ pi: a reply over the cap is a tool error naming the size and the cap; list_tasks asks for open tasks (310.922905ms)
✔ pi: a write through a dangling symlink is blocked, and nothing appears outside (356.448684ms)
✔ pi: a read is refused when pi would open another spelling outside (313.987297ms)
✔ pi: a relative path climbs from the real path of a workspace behind a symlink (309.279237ms)
✔ pi: a relative path climbs from the real path of a dataRoot behind a symlink (299.482402ms)
✔ pi: a missing extension refuses before any model call (6.408932ms)
✔ pi: an extension without its configuration fails pi's start (242.65042ms)
✔ founderCheck: founder variables, then a needed service without a usable token (1.387414ms)
✔ turnRequest names the sender, class, reply and decision (0.231666ms)
✔ turnRequest says a tester's message is a test and never Jason's instruction (0.191222ms)
✔ turnRequest shows the message's own sent time, a day old as well as new (0.140888ms)
✔ stateBlock names the launch state with its time and event, and the open tasks with the cap (0.231872ms)
✔ a message becomes a turn, the answer goes back as a RESULT, SIGTERM releases and exits 0 (211.037409ms)
✔ launches revoked and then restored: the turn's block reads allowed and gives the restore time (168.641165ms)
✔ a task closed on the tracker is missing from the next turn's open list (168.874887ms)
✔ a message read a day after it was sent shows the time it was sent (88.493867ms)
✔ a decision the human resolves becomes a turn citing the decision and the answer (167.958612ms)
✔ a failed state read fails the turn: no adapter runs, the sender hears why, the message is read (64.332042ms)
✔ a SIGTERM before the claim stops the runner with exit 0 and no claim (48.971564ms)
✔ typed tools carry the runner's capability; launch goes to the host's launch socket (272.7423ms)
✔ a RESULT gets no automatic reply; failed turns reply with the reason (1197.197203ms)
✔ SIGTERM during a turn kills the turn's process group and still exits 0 (92.595515ms)
✔ founder credentials stop before the claim (20) (96.308192ms)
✔ a refused claim exits 21; an ended run's capability exits 22 (95.162912ms)
✔ the launch ending under a running session exits 22 (99.199269ms)
✔ a broker that stays unreachable exits 23 after brokerRetries polls (204.764604ms)
✔ a broker that is down at the claim exits 23, not 21 (50.652923ms)
✔ no capability, or a malformed one, on stdin exits 2 (142.013049ms)
✔ a missing or malformed policy exits 2 before the claim (85.249607ms)
✔ the PM gets its task verbs and the reads, and no launch (7.157617ms)
✔ the chief of staff gets launch and send_message, and no task verbs (2.836501ms)
✔ the assistant gets send_message and the reads, no launch (2.833563ms)
✔ list_requirements reads the requirement ids, for a task.create holder with a tracker (3.577265ms)
✔ a coder gets no launch, no resolve_decision, and no task tools without a tracker (1.237907ms)
✔ a coder with a tracker gets its own task tools and no task_reassign (1.569331ms)
✔ launch only when the business's launch block names the instance as launcher (2.600909ms)
✔ an action outside the instance's authority has no tool (1.452998ms)
✔ the review loop: push_working and request_review for the coder, fetch_review and review_verdict for the reviewer, none for the PM (4.485505ms)
✔ callTool: one JSON line out, the result back, a refusal rejects (11.338377ms)
✔ list_tasks takes the broker's filters and pages; get_task takes a task_ref (3.681892ms)
✔ callTool sends list_tasks done false unless the session names done; other tools as given (3.583033ms)
✔ replyText: compact JSON up to the cap, an error naming the size and the cap over it (1.246407ms)
ℹ tests 84
ℹ suites 0
ℹ pass 84
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 10606.422363
