✔ sessionModel: agent vars win, then the system's execution settings (15.071484ms) ✔ a pi bundle: prompt, policy, tools and manifest, 0600 in a 0700 directory (4.702614ms) ✔ a chief-of-staff bundle lists launch among the bus tools and the manifest's typed tools (3.42425ms) ✔ a chief-of-staff bundle lists end_run and change_organization as typed and nothing untyped, for either harness (4.420311ms) ✔ the stop line names end_run for a role.end holder, and every other role keeps today's text (7.625173ms) ✔ a claude-code bundle adds the wrapped gate hook and the MCP config (2.058729ms) ✔ the prompt's built-in tools line is the harness's own names, as the manifest lists them (8.831773ms) ✔ the prompt names the project clones in the workspace, and none when there are none (5.681984ms) ✔ the prompt explains the review loop to a role with push_working, and reviews to one with fetch_review (5.455545ms) ✔ a bundle is written once: an existing file refuses (1.593387ms) ✔ a path with a single quote can't go into the hook command (1.556315ms) ✔ allow exits 0, a deny exits 2 with the reason on stderr (130.107933ms) ✔ a missing or wrong policy, or a bad event, exits 2 (82.316236ms) ✔ the bundle's wrapped command: a missing gate or node still blocks (1090.252686ms) ✔ claude: typed tools through MCP, the hook blocks, builtins outside --tools don't exist (739.103324ms) ✔ claude: the hook alone blocks a path outside the workspace (434.769991ms) ✔ claude: a second turn resumes the first turn's session (729.319385ms) ✔ claude adapter: --restricted is always passed (5.024244ms) ✔ claude adapter: MOSAIC_CLAUDE_BIN wins over PATH and must be an absolute executable (15.738382ms) ✔ claude adapter: a directory pin exits 2 before it runs; a pin with a space runs (5.719471ms) ✔ claude: CLAUDE.md files and auto-memory don't reach the model; without --restricted they do (723.817314ms) ✔ claude: a missing hook or MCP file refuses before claude starts (7.362201ms) ✔ pi: policy tools and typed tools pass, anything else is blocked (4.24262ms) ✔ claude: builtins map from pi names, typed tools need the mcp prefix (0.985612ms) ✔ file tool paths must resolve inside the workspace (1.597691ms) ✔ pi's own path normalisation can't be used to step out (0.900919ms) ✔ a symlink inside the workspace that points out is outside (0.802971ms) ✔ a dangling symlink is refused at any depth, in both harnesses (2.717461ms) ✔ read is checked under every spelling pi's read would open, in both harnesses (14.050884ms) ✔ other spellings cover directories, dangling links and pi's cwd (2.304323ms) ✔ a relative path climbs from the workspace's real path, in both harnesses (6.901009ms) ✔ claude path fields per tool (0.804117ms) ✔ glob patterns stay inside the workspace (0.696706ms) ✔ a path that can't be checked is blocked (0.402181ms) ✔ fetch_review's have lists the working refs before origin's, and no more than HAVE_MAX (211.488985ms) ✔ fetch_review's have orders the working refs newest first within each clone, one clone after another (134.735046ms) ✔ a runner git past its time: SIGTERM reaches its process group, and SIGKILL follows after the grace period (1809.637545ms) ✔ once a signalled runner git has closed, its group gets SIGKILL, so a process that ignored SIGTERM dies too (303.941413ms) ✔ initialize, ping and tools/list (42.930529ms) ✔ tools/call goes through the tool socket; a refusal is an isError result (33.645613ms) ✔ replies are compact JSON; over the cap a tool error names the size and the cap (35.89632ms) ✔ unknown tools and methods are JSON-RPC errors and never reach the socket (30.031364ms) ✔ a missing argument is a usage error (26.161545ms) ✔ pi: typed tools reach the socket, the gate blocks, agent_end writes the marker (356.656124ms) ✔ pi: a reply over the cap is a tool error naming the size and the cap; list_tasks asks for open tasks (307.923138ms) ✔ pi: a write through a dangling symlink is blocked, and nothing appears outside (299.233526ms) ✔ pi: a read is refused when pi would open another spelling outside (313.192805ms) ✔ pi: a relative path climbs from the real path of a workspace behind a symlink (303.958691ms) ✔ pi: a relative path climbs from the real path of a dataRoot behind a symlink (309.869099ms) ✔ pi: a missing extension refuses before any model call (5.620109ms) ✔ pi: an extension without its configuration fails pi's start (247.357276ms) ✔ founderCheck: founder variables, then a needed service without a usable token (1.785408ms) ✔ turnRequest names the sender, class, reply and decision (0.302386ms) ✔ turnRequest says a tester's message is a test and never Jason's instruction (0.271857ms) ✔ turnRequest shows the message's own sent time, a day old as well as new (0.184337ms) ✔ stateBlock names the launch state with its time and event, and the open tasks with the cap (0.26359ms) ✔ a message becomes a turn, the answer goes back as a RESULT, SIGTERM releases and exits 0 (266.394661ms) ✔ launches revoked and then restored: the turn's block reads allowed and gives the restore time (247.494639ms) ✔ a task closed on the tracker is missing from the next turn's open list (219.255713ms) ✔ a message read a day after it was sent shows the time it was sent (169.986685ms) ✔ a decision the human resolves becomes a turn citing the decision and the answer (207.525124ms) ✔ a failed state read fails the turn: no adapter runs, the sender hears why, the message is read (123.588329ms) ✔ a SIGTERM before the claim stops the runner with exit 0 and no claim (119.158729ms) ✔ typed tools carry the runner's capability; launch goes to the host's launch socket (396.391687ms) ✔ end_run and change_organization reach the broker with the tool's arguments, and refusals keep their codes (469.362511ms) ✔ a RESULT gets no automatic reply; failed turns reply with the reason (1288.315914ms) ✔ SIGTERM during a turn kills the turn's process group and still exits 0 (149.543957ms) ✔ founder credentials stop before the claim (20) (159.751317ms) ✔ a refused claim exits 21; an ended run's capability exits 22 (183.101955ms) ✔ the launch ending under a running session exits 22 (142.031359ms) ✔ a broker that stays unreachable exits 23 after brokerRetries polls (242.529605ms) ✔ a broker that is down at the claim exits 23, not 21 (94.356119ms) ✔ no capability, or a malformed one, on stdin exits 2 (193.184282ms) ✔ a missing or malformed policy exits 2 before the claim (130.541434ms) ✔ the PM gets its task verbs and the reads, and no launch (7.269696ms) ✔ the chief of staff gets launch, send_message, end_run and change_organization, and no task verbs (2.370533ms) ✔ end_run sends role and reason, no decision; change_organization sends revision and the organization keys (2.650787ms) ✔ the organization keys change_organization takes are the broker's ORGANIZATION_KEYS (0.180088ms) ✔ a role without role.end or organization.change gets neither tool, and each tool follows its own grant (10.053174ms) ✔ the assistant gets send_message and the reads, no launch (1.631192ms) ✔ list_requirements reads the requirement ids, for a task.create holder with a tracker (2.169698ms) ✔ a coder gets no launch, no resolve_decision, and no task tools without a tracker (1.026766ms) ✔ a coder with a tracker gets its own task tools and no task_reassign (1.273896ms) ✔ launch only when the business's launch block names the instance as launcher (1.867198ms) ✔ an action outside the instance's authority has no tool (1.066824ms) ✔ the review loop: push_working and request_review for the coder, fetch_review and review_verdict for the reviewer, none for the PM (5.142215ms) ✔ callTool: one JSON line out, the result back, a refusal rejects (13.20392ms) ✔ list_tasks takes the broker's filters and pages; get_task takes a task_ref (3.251887ms) ✔ callTool sends list_tasks done false unless the session names done; other tools as given (3.364973ms) ✔ replyText: compact JSON up to the cap, an error naming the size and the cap over it (0.87286ms) ℹ tests 90 ℹ suites 0 ℹ pass 90 ℹ fail 0 ℹ cancelled 0 ℹ skipped 0 ℹ todo 0 ℹ duration_ms 11789.303058