✔ sessionModel: agent vars win, then the system's execution settings (12.889085ms) ✔ a pi bundle: prompt, policy, tools and manifest, 0600 in a 0700 directory (5.085926ms) ✔ a chief-of-staff bundle lists launch among the bus tools and the manifest's typed tools (2.223859ms) ✔ a claude-code bundle adds the wrapped gate hook and the MCP config (2.257ms) ✔ the prompt's built-in tools line is the harness's own names, as the manifest lists them (8.745906ms) ✔ the prompt names the project clones in the workspace, and none when there are none (6.689266ms) ✔ the prompt explains the review loop to a role with push_working, and reviews to one with fetch_review (7.586157ms) ✔ a bundle is written once: an existing file refuses (1.436797ms) ✔ a path with a single quote can't go into the hook command (2.026647ms) ✔ allow exits 0, a deny exits 2 with the reason on stderr (127.564329ms) ✔ a missing or wrong policy, or a bad event, exits 2 (90.199177ms) ✔ the bundle's wrapped command: a missing gate or node still blocks (1080.268615ms) ✔ claude: typed tools through MCP, the hook blocks, builtins outside --tools don't exist (746.417629ms) ✔ claude: the hook alone blocks a path outside the workspace (429.797634ms) ✔ claude: a second turn resumes the first turn's session (739.201231ms) ✔ claude adapter: --restricted is always passed (5.31313ms) ✔ claude adapter: MOSAIC_CLAUDE_BIN wins over PATH and must be an absolute executable (16.640679ms) ✔ claude adapter: a directory pin exits 2 before it runs; a pin with a space runs (5.833519ms) ✔ claude: CLAUDE.md files and auto-memory don't reach the model; without --restricted they do (703.469709ms) ✔ claude: a missing hook or MCP file refuses before claude starts (7.232319ms) ✔ pi: policy tools and typed tools pass, anything else is blocked (2.894013ms) ✔ claude: builtins map from pi names, typed tools need the mcp prefix (0.601348ms) ✔ file tool paths must resolve inside the workspace (1.174216ms) ✔ pi's own path normalisation can't be used to step out (0.940164ms) ✔ a symlink inside the workspace that points out is outside (0.84133ms) ✔ a dangling symlink is refused at any depth, in both harnesses (3.299988ms) ✔ read is checked under every spelling pi's read would open, in both harnesses (16.749325ms) ✔ other spellings cover directories, dangling links and pi's cwd (1.473866ms) ✔ a relative path climbs from the workspace's real path, in both harnesses (4.689979ms) ✔ claude path fields per tool (0.920298ms) ✔ glob patterns stay inside the workspace (0.905666ms) ✔ a path that can't be checked is blocked (0.370846ms) ✔ fetch_review's have lists the working refs before origin's, and no more than HAVE_MAX (214.936069ms) ✔ fetch_review's have orders the working refs newest first within each clone, one clone after another (127.758917ms) ✔ a runner git past its time: SIGTERM reaches its process group, and SIGKILL follows after the grace period (1809.740168ms) ✔ once a signalled runner git has closed, its group gets SIGKILL, so a process that ignored SIGTERM dies too (303.908478ms) ✔ initialize, ping and tools/list (40.728252ms) ✔ tools/call goes through the tool socket; a refusal is an isError result (34.156428ms) ✔ replies are compact JSON; over the cap a tool error names the size and the cap (40.960811ms) ✔ unknown tools and methods are JSON-RPC errors and never reach the socket (29.155409ms) ✔ a missing argument is a usage error (26.566392ms) ✔ pi: typed tools reach the socket, the gate blocks, agent_end writes the marker (365.298149ms) ✔ pi: a reply over the cap is a tool error naming the size and the cap; list_tasks asks for open tasks (316.991495ms) ✔ pi: a write through a dangling symlink is blocked, and nothing appears outside (306.660399ms) ✔ pi: a read is refused when pi would open another spelling outside (311.979704ms) ✔ pi: a relative path climbs from the real path of a workspace behind a symlink (313.416635ms) ✔ pi: a relative path climbs from the real path of a dataRoot behind a symlink (314.089216ms) ✔ pi: a missing extension refuses before any model call (5.77622ms) ✔ pi: an extension without its configuration fails pi's start (256.662603ms) ✔ founderCheck: founder variables, then a needed service without a usable token (1.387046ms) ✔ turnRequest names the sender, class, reply and decision (0.253092ms) ✔ turnRequest says a tester's message is a test and never Jason's instruction (0.211573ms) ✔ turnRequest shows the message's own sent time, a day old as well as new (0.164465ms) ✔ stateBlock names the launch state with its time and event, and the open tasks with the cap (0.258464ms) ✔ a message becomes a turn, the answer goes back as a RESULT, SIGTERM releases and exits 0 (252.24083ms) ✔ launches revoked and then restored: the turn's block reads allowed and gives the restore time (258.899685ms) ✔ a task closed on the tracker is missing from the next turn's open list (217.793193ms) ✔ a message read a day after it was sent shows the time it was sent (165.764663ms) ✔ a decision the human resolves becomes a turn citing the decision and the answer (203.540823ms) ✔ a failed state read fails the turn: no adapter runs, the sender hears why, the message is read (126.659738ms) ✔ a SIGTERM before the claim stops the runner with exit 0 and no claim (105.248413ms) ✔ typed tools carry the runner's capability; launch goes to the host's launch socket (400.893449ms) ✔ a RESULT gets no automatic reply; failed turns reply with the reason (1293.861118ms) ✔ SIGTERM during a turn kills the turn's process group and still exits 0 (152.987087ms) ✔ founder credentials stop before the claim (20) (170.957294ms) ✔ a refused claim exits 21; an ended run's capability exits 22 (190.611433ms) ✔ the launch ending under a running session exits 22 (144.926994ms) ✔ a broker that stays unreachable exits 23 after brokerRetries polls (242.343432ms) ✔ a broker that is down at the claim exits 23, not 21 (89.137101ms) ✔ no capability, or a malformed one, on stdin exits 2 (179.299653ms) ✔ a missing or malformed policy exits 2 before the claim (122.386806ms) ✔ the PM gets its task verbs and the reads, and no launch (8.104718ms) ✔ the chief of staff gets launch and send_message, and no task verbs (2.239896ms) ✔ the assistant gets send_message and the reads, no launch (2.696098ms) ✔ list_requirements reads the requirement ids, for a task.create holder with a tracker (3.171359ms) ✔ a coder gets no launch, no resolve_decision, and no task tools without a tracker (1.355345ms) ✔ a coder with a tracker gets its own task tools and no task_reassign (1.308194ms) ✔ launch only when the business's launch block names the instance as launcher (2.770474ms) ✔ an action outside the instance's authority has no tool (1.507608ms) ✔ the review loop: push_working and request_review for the coder, fetch_review and review_verdict for the reviewer, none for the PM (4.043643ms) ✔ callTool: one JSON line out, the result back, a refusal rejects (8.020255ms) ✔ list_tasks takes the broker's filters and pages; get_task takes a task_ref (2.461054ms) ✔ callTool sends list_tasks done false unless the session names done; other tools as given (2.653016ms) ✔ replyText: compact JSON up to the cap, an error naming the size and the cap over it (0.798163ms) ℹ tests 84 ℹ suites 0 ℹ pass 84 ℹ fail 0 ℹ cancelled 0 ℹ skipped 0 ℹ todo 0 ℹ duration_ms 11299.273443