✔ sessionModel: agent vars win, then the system's execution settings (14.690212ms) ✔ a pi bundle: prompt, policy, tools and manifest, 0600 in a 0700 directory (6.449921ms) ✔ a chief-of-staff bundle lists launch among the bus tools and the manifest's typed tools (3.714709ms) ✔ a chief-of-staff bundle lists end_run and change_organization as typed and only instance.name untyped, for either harness (6.951046ms) ✔ the stop line names end_run for a role.end holder, and every other role keeps today's text (6.889571ms) ✔ a claude-code bundle adds the wrapped gate hook and the MCP config (3.21908ms) ✔ the prompt's built-in tools line is the harness's own names, as the manifest lists them (13.978955ms) ✔ the prompt names the project clones in the workspace, and none when there are none (8.575428ms) ✔ the prompt explains the review loop to a role with push_working, and reviews to one with fetch_review (13.680696ms) ✔ the prompt explains landing to a role with land_work, and to no other (7.479532ms) ✔ the prompt tells the pm to ask the chief of staff for sessions, and no other role (11.355204ms) ✔ a bundle is written once: an existing file refuses (3.496005ms) ✔ a path with a single quote can't go into the hook command (2.23778ms) ✔ web grant on pi: the bundle has web_search and web_fetch, and no built-in web tool (2.775334ms) ✔ no web grant on pi: no web tool in the bundle, and the gate blocks a web call (2.568326ms) ✔ web grant on claude-code: the bundle has web_search and web_fetch, and no built-in web tool (2.617767ms) ✔ no web grant on claude-code: no web tool in the bundle, and the gate blocks a web call (4.468307ms) ✔ a granted session's gate lets the typed web tools through and still blocks the built-ins (2.610941ms) ✔ search-source searxng-url: a granted role gets the typed broker tools and no harness web tool; an ungranted role gets nothing (7.807654ms) ✔ search-source searxng-bundled: a granted role gets the typed broker tools and no harness web tool; an ungranted role gets nothing (8.414105ms) ✔ search-source harness-builtin: claude-code gets WebSearch and WebFetch and no typed web tool, only when granted (11.711529ms) ✔ search-source harness-builtin is refused for a granted pi role, which has no built-in web tool (2.856373ms) ✔ allow exits 0, a deny exits 2 with the reason on stderr (193.729318ms) ✔ a missing or wrong policy, or a bad event, exits 2 (89.303527ms) ✔ the bundle's wrapped command: a missing gate or node still blocks (1089.299557ms) ✔ claude: typed tools through MCP, the hook blocks, builtins outside --tools don't exist (841.232899ms) ✔ claude: the hook alone blocks a path outside the workspace (448.354761ms) ✔ claude: a second turn resumes the first turn's session (739.454972ms) ✔ claude adapter: --restricted is always passed (5.869829ms) ✔ claude adapter: MOSAIC_CLAUDE_BIN wins over PATH and must be an absolute executable (17.09829ms) ✔ claude adapter: a directory pin exits 2 before it runs; a pin with a space runs (5.545202ms) ✔ claude: CLAUDE.md files and auto-memory don't reach the model; without --restricted they do (740.939966ms) ✔ claude: a missing hook or MCP file refuses before claude starts (7.49754ms) ✔ pi: policy tools and typed tools pass, anything else is blocked (4.319148ms) ✔ claude: builtins map from pi names, typed tools need the mcp prefix (0.979954ms) ✔ file tool paths must resolve inside the workspace (1.774166ms) ✔ pi's own path normalisation can't be used to step out (1.364441ms) ✔ a symlink inside the workspace that points out is outside (1.163269ms) ✔ a dangling symlink is refused at any depth, in both harnesses (3.78995ms) ✔ read is checked under every spelling pi's read would open, in both harnesses (32.758434ms) ✔ other spellings cover directories, dangling links and pi's cwd (3.989415ms) ✔ a relative path climbs from the workspace's real path, in both harnesses (8.246192ms) ✔ claude path fields per tool (1.264599ms) ✔ glob patterns stay inside the workspace (1.075516ms) ✔ a path that can't be checked is blocked (0.665237ms) ✔ fetch_review's have lists the working refs before origin's, and no more than HAVE_MAX (300.499956ms) ✔ fetch_review's have orders the working refs newest first within each clone, one clone after another (135.112706ms) ✔ a runner git past its time: SIGTERM reaches its process group, and SIGKILL follows after the grace period (1810.00311ms) ✔ once a signalled runner git has closed, its group gets SIGKILL, so a process that ignored SIGTERM dies too (303.843061ms) ✔ land_work sends one land line and resolves the host's started answer (3.138729ms) ✔ initialize, ping and tools/list (53.872046ms) ✔ tools/call goes through the tool socket; a refusal is an isError result (61.292158ms) ✔ replies are compact JSON; over the cap a tool error names the size and the cap (57.128624ms) ✔ unknown tools and methods are JSON-RPC errors and never reach the socket (37.749133ms) ✔ a missing argument is a usage error (34.038403ms) ✔ pi: typed tools reach the socket, the gate blocks, agent_end writes the marker (426.865751ms) ✔ pi: a reply over the cap is a tool error naming the size and the cap; list_tasks asks for open tasks (307.182309ms) ✔ pi: a write through a dangling symlink is blocked, and nothing appears outside (309.761202ms) ✔ pi: a read is refused when pi would open another spelling outside (310.236977ms) ✔ pi: a relative path climbs from the real path of a workspace behind a symlink (299.589299ms) ✔ pi: a relative path climbs from the real path of a dataRoot behind a symlink (301.302294ms) ✔ pi: a missing extension refuses before any model call (6.186471ms) ✔ pi: an extension without its configuration fails pi's start (254.232023ms) ✔ founderCheck: founder variables, then a needed service without a usable token (1.149366ms) ✔ turnRequest names the sender, class, reply and decision (0.220155ms) ✔ turnRequest says a tester's message is a test and never Jason's instruction (0.234894ms) ✔ turnRequest shows the message's own sent time, a day old as well as new (0.134208ms) ✔ stateBlock names the launch state with its time and event, and the open tasks with the cap (0.220732ms) ✔ a message becomes a turn, the answer goes back as a RESULT, SIGTERM releases and exits 0 (335.916658ms) ✔ launches revoked and then restored: the turn's block reads allowed and gives the restore time (272.764119ms) ✔ a task closed on the tracker is missing from the next turn's open list (235.899577ms) ✔ a message read a day after it was sent shows the time it was sent (182.538405ms) ✔ a decision the human resolves becomes a turn citing the decision and the answer (231.558095ms) ✔ a failed state read fails the turn: no adapter runs, the sender hears why, the message is read (170.742775ms) ✔ a SIGTERM before the claim stops the runner with exit 0 and no claim (105.371778ms) ✔ typed tools carry the runner's capability; launch goes to the host's launch socket (399.941144ms) ✔ end_run and change_organization reach the broker with the tool's arguments, and refusals keep their codes (463.589077ms) ✔ a RESULT gets no automatic reply; failed turns reply with the reason (1298.204829ms) ✔ the bus's launch notice is one turn with no reply, and is not taken again (286.588214ms) ✔ SIGTERM during a turn kills the turn's process group and still exits 0 (156.112799ms) ✔ founder credentials stop before the claim (20) (183.328607ms) ✔ a refused claim exits 21; an ended run's capability exits 22 (212.57649ms) ✔ the launch ending under a running session exits 22 (154.4556ms) ✔ a broker that stays unreachable exits 23 after brokerRetries polls (251.153593ms) ✔ a broker that is down at the claim exits 23, not 21 (95.656923ms) ✔ no capability, or a malformed one, on stdin exits 2 (198.192759ms) ✔ a missing or malformed policy exits 2 before the claim (132.772121ms) ✔ every role sees read_spec, record writer follows its grant, citation schemas stay strict (22.226228ms) ✔ typed socket forwards read content/history and agent record writes, and cannot write a spec (242.476158ms) ✔ the PM gets its task verbs and the reads, and no launch (8.444087ms) ✔ send_message offers the BLOCK class, beside the classes it already had (3.134543ms) ✔ the chief of staff gets launch, send_message, end_run and change_organization, and no task verbs (2.425062ms) ✔ request_staff is the pm's alone: staff.request over message.send, roles and counts bounded (14.338126ms) ✔ end_run sends role and reason, no decision; change_organization sends revision and the organization keys (2.121036ms) ✔ the organization keys change_organization takes are the broker's ORGANIZATION_KEYS (0.155018ms) ✔ a role without role.end or organization.change gets neither tool, and each tool follows its own grant (15.002005ms) ✔ the assistant gets send_message, the reads and the web tools (it holds the web grant), no launch (3.701218ms) ✔ list_requirements reads the requirement ids, for a task.create holder with a tracker (5.94431ms) ✔ a coder gets no launch, no resolve_decision, and no task tools without a tracker (1.694162ms) ✔ a coder with a tracker gets its own task tools and no task_reassign (1.70071ms) ✔ launch only when the business's launch block names the instance as launcher (3.263894ms) ✔ an action outside the instance's authority has no tool (1.857942ms) ✔ the review loop: push_working and request_review for the coder, fetch_review and review_verdict for the reviewer, none for the PM (7.16009ms) ✔ land_work is there only with land, after push_working, with the task and the decision (4.288067ms) ✔ callTool: one JSON line out, the result back, a refusal rejects (17.310182ms) ✔ list_tasks takes the broker's filters and pages; get_task takes a task_ref (4.265729ms) ✔ callTool sends list_tasks done false unless the session names done; other tools as given (5.959713ms) ✔ replyText: compact JSON up to the cap, an error naming the size and the cap over it (1.011099ms) ✔ web_search's query maxLength is the web layer's SEARCH_MAX_QUERY_CHARS (5.057892ms) ℹ tests 109 ℹ suites 0 ℹ pass 109 ℹ fail 0 ℹ cancelled 0 ℹ skipped 0 ℹ todo 0 ℹ duration_ms 12269.2165