✔ sessionModel: agent vars win, then the system's execution settings (11.343424ms)
✔ a pi bundle: prompt, policy, tools and manifest, 0600 in a 0700 directory (5.782665ms)
✔ a chief-of-staff bundle lists launch among the bus tools and the manifest's typed tools (3.43918ms)
✔ a chief-of-staff bundle lists end_run and change_organization as typed and nothing untyped, for either harness (5.933102ms)
✔ the stop line names end_run for a role.end holder, and every other role keeps today's text (6.749438ms)
✔ a claude-code bundle adds the wrapped gate hook and the MCP config (2.069824ms)
✔ the prompt's built-in tools line is the harness's own names, as the manifest lists them (8.481968ms)
✔ the prompt names the project clones in the workspace, and none when there are none (6.860294ms)
✔ the prompt explains the review loop to a role with push_working, and reviews to one with fetch_review (6.289839ms)
✔ the prompt explains landing to a role with land_work, and to no other (4.388041ms)
✔ the prompt tells the pm to ask the chief of staff for sessions, and no other role (8.920358ms)
✔ a bundle is written once: an existing file refuses (1.987774ms)
✔ a path with a single quote can't go into the hook command (1.956159ms)
✔ web grant on pi: the bundle has web_search and web_fetch, and no built-in web tool (1.995691ms)
✔ no web grant on pi: no web tool in the bundle, and the gate blocks a web call (1.785105ms)
✔ web grant on claude-code: the bundle has web_search and web_fetch, and no built-in web tool (2.067121ms)
✔ no web grant on claude-code: no web tool in the bundle, and the gate blocks a web call (1.560598ms)
✔ a granted session's gate lets the typed web tools through and still blocks the built-ins (1.548046ms)
✔ search-source searxng-url: a granted role gets the typed broker tools and no harness web tool; an ungranted role gets nothing (5.288361ms)
✔ search-source searxng-bundled: a granted role gets the typed broker tools and no harness web tool; an ungranted role gets nothing (6.534577ms)
✔ search-source harness-builtin: claude-code gets WebSearch and WebFetch and no typed web tool, only when granted (6.493394ms)
✔ search-source harness-builtin is refused for a granted pi role, which has no built-in web tool (1.694055ms)
✔ allow exits 0, a deny exits 2 with the reason on stderr (141.081577ms)
✔ a missing or wrong policy, or a bad event, exits 2 (80.049122ms)
✔ the bundle's wrapped command: a missing gate or node still blocks (1091.009456ms)
✔ claude: typed tools through MCP, the hook blocks, builtins outside --tools don't exist (738.45514ms)
✔ claude: the hook alone blocks a path outside the workspace (410.328058ms)
✔ claude: a second turn resumes the first turn's session (692.221133ms)
✔ claude adapter: --restricted is always passed (5.336195ms)
✔ claude adapter: MOSAIC_CLAUDE_BIN wins over PATH and must be an absolute executable (16.227306ms)
✔ claude adapter: a directory pin exits 2 before it runs; a pin with a space runs (5.448558ms)
✔ claude: CLAUDE.md files and auto-memory don't reach the model; without --restricted they do (664.489857ms)
✔ claude: a missing hook or MCP file refuses before claude starts (7.298151ms)
✔ pi: policy tools and typed tools pass, anything else is blocked (4.297496ms)
✔ claude: builtins map from pi names, typed tools need the mcp prefix (1.028537ms)
✔ file tool paths must resolve inside the workspace (1.648287ms)
✔ pi's own path normalisation can't be used to step out (1.347879ms)
✔ a symlink inside the workspace that points out is outside (1.191131ms)
✔ a dangling symlink is refused at any depth, in both harnesses (4.167715ms)
✔ read is checked under every spelling pi's read would open, in both harnesses (16.783878ms)
✔ other spellings cover directories, dangling links and pi's cwd (1.528809ms)
✔ a relative path climbs from the workspace's real path, in both harnesses (5.810483ms)
✔ claude path fields per tool (0.992895ms)
✔ glob patterns stay inside the workspace (0.862447ms)
✔ a path that can't be checked is blocked (0.742127ms)
✔ fetch_review's have lists the working refs before origin's, and no more than HAVE_MAX (220.331472ms)
✔ fetch_review's have orders the working refs newest first within each clone, one clone after another (142.517683ms)
✔ a runner git past its time: SIGTERM reaches its process group, and SIGKILL follows after the grace period (1809.709817ms)
✔ once a signalled runner git has closed, its group gets SIGKILL, so a process that ignored SIGTERM dies too (304.160591ms)
✔ land_work sends one land line and resolves the host's started answer (3.205962ms)
✔ initialize, ping and tools/list (48.147763ms)
✔ tools/call goes through the tool socket; a refusal is an isError result (47.792708ms)
✔ replies are compact JSON; over the cap a tool error names the size and the cap (37.000089ms)
✔ unknown tools and methods are JSON-RPC errors and never reach the socket (31.714003ms)
✔ a missing argument is a usage error (24.267472ms)
✔ pi: typed tools reach the socket, the gate blocks, agent_end writes the marker (378.934894ms)
✔ pi: a reply over the cap is a tool error naming the size and the cap; list_tasks asks for open tasks (309.077715ms)
✔ pi: a write through a dangling symlink is blocked, and nothing appears outside (325.285952ms)
✔ pi: a read is refused when pi would open another spelling outside (303.642451ms)
✔ pi: a relative path climbs from the real path of a workspace behind a symlink (326.359659ms)
✔ pi: a relative path climbs from the real path of a dataRoot behind a symlink (294.019831ms)
✔ pi: a missing extension refuses before any model call (5.896337ms)
✔ pi: an extension without its configuration fails pi's start (245.965663ms)
✔ founderCheck: founder variables, then a needed service without a usable token (1.141201ms)
✔ turnRequest names the sender, class, reply and decision (0.279464ms)
✔ turnRequest says a tester's message is a test and never Jason's instruction (0.211734ms)
✔ turnRequest shows the message's own sent time, a day old as well as new (0.12921ms)
✔ stateBlock names the launch state with its time and event, and the open tasks with the cap (0.244861ms)
✔ a message becomes a turn, the answer goes back as a RESULT, SIGTERM releases and exits 0 (190.677466ms)
✔ launches revoked and then restored: the turn's block reads allowed and gives the restore time (176.149697ms)
✔ a task closed on the tracker is missing from the next turn's open list (170.595684ms)
✔ a message read a day after it was sent shows the time it was sent (89.037839ms)
✔ a decision the human resolves becomes a turn citing the decision and the answer (141.745596ms)
✔ a failed state read fails the turn: no adapter runs, the sender hears why, the message is read (64.040043ms)
✔ a SIGTERM before the claim stops the runner with exit 0 and no claim (50.99337ms)
✔ typed tools carry the runner's capability; launch goes to the host's launch socket (274.46281ms)
✔ end_run and change_organization reach the broker with the tool's arguments, and refusals keep their codes (283.909952ms)
✔ a RESULT gets no automatic reply; failed turns reply with the reason (1195.164234ms)
✔ the bus's launch notice is one turn with no reply, and is not taken again (217.40609ms)
✔ SIGTERM during a turn kills the turn's process group and still exits 0 (94.940596ms)
✔ founder credentials stop before the claim (20) (90.187982ms)
✔ a refused claim exits 21; an ended run's capability exits 22 (95.42765ms)
✔ the launch ending under a running session exits 22 (98.896296ms)
✔ a broker that stays unreachable exits 23 after brokerRetries polls (198.689225ms)
✔ a broker that is down at the claim exits 23, not 21 (48.723128ms)
✔ no capability, or a malformed one, on stdin exits 2 (148.994173ms)
✔ a missing or malformed policy exits 2 before the claim (82.519493ms)
✔ every role sees read_spec, record writer follows its grant, citation schemas stay strict (17.027167ms)
✔ typed socket forwards read content/history and agent record writes, and cannot write a spec (97.785823ms)
✔ the PM gets its task verbs and the reads, and no launch (8.920022ms)
✔ the chief of staff gets launch, send_message, end_run and change_organization, and no task verbs (4.33593ms)
✔ request_staff is the pm's alone: staff.request over message.send, roles and counts bounded (10.284732ms)
✔ end_run sends role and reason, no decision; change_organization sends revision and the organization keys (1.873738ms)
✔ the organization keys change_organization takes are the broker's ORGANIZATION_KEYS (0.149279ms)
✔ a role without role.end or organization.change gets neither tool, and each tool follows its own grant (7.948528ms)
✔ the assistant gets send_message, the reads and the web tools (it holds the web grant), no launch (1.268932ms)
✔ list_requirements reads the requirement ids, for a task.create holder with a tracker (2.503733ms)
✔ a coder gets no launch, no resolve_decision, and no task tools without a tracker (0.942871ms)
✔ a coder with a tracker gets its own task tools and no task_reassign (0.917828ms)
✔ launch only when the business's launch block names the instance as launcher (1.870071ms)
✔ an action outside the instance's authority has no tool (1.31595ms)
✔ the review loop: push_working and request_review for the coder, fetch_review and review_verdict for the reviewer, none for the PM (4.424238ms)
✔ land_work is there only with land, after push_working, with the task and the decision (2.306375ms)
✔ callTool: one JSON line out, the result back, a refusal rejects (11.56548ms)
✔ list_tasks takes the broker's filters and pages; get_task takes a task_ref (2.956308ms)
✔ callTool sends list_tasks done false unless the session names done; other tools as given (2.582422ms)
✔ replyText: compact JSON up to the cap, an error naming the size and the cap over it (0.720384ms)
✔ web_search's query maxLength is the web layer's SEARCH_MAX_QUERY_CHARS (3.901222ms)
ℹ tests 108
ℹ suites 0
ℹ pass 108
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 11099.092638
