✔ sessionModel: agent vars win, then the system's execution settings (15.426201ms) ✔ a pi bundle: prompt, policy, tools and manifest, 0600 in a 0700 directory (5.953766ms) ✔ a chief-of-staff bundle lists launch among the bus tools and the manifest's typed tools (3.970248ms) ✔ a chief-of-staff bundle lists end_run and change_organization as typed and nothing untyped, for either harness (5.896742ms) ✔ the stop line names end_run for a role.end holder, and every other role keeps today's text (10.174631ms) ✔ a claude-code bundle adds the wrapped gate hook and the MCP config (2.805548ms) ✔ the prompt's built-in tools line is the harness's own names, as the manifest lists them (8.604071ms) ✔ the prompt names the project clones in the workspace, and none when there are none (5.986235ms) ✔ the prompt explains the review loop to a role with push_working, and reviews to one with fetch_review (6.387195ms) ✔ the prompt explains landing to a role with land_work, and to no other (3.738884ms) ✔ the prompt tells the pm to ask the chief of staff for sessions, and no other role (6.416866ms) ✔ a bundle is written once: an existing file refuses (1.314855ms) ✔ a path with a single quote can't go into the hook command (1.269205ms) ✔ web grant on pi: the bundle has web_search and web_fetch, and no built-in web tool (2.473563ms) ✔ no web grant on pi: no web tool in the bundle, and the gate blocks a web call (1.666416ms) ✔ web grant on claude-code: the bundle has web_search and web_fetch, and no built-in web tool (1.515497ms) ✔ no web grant on claude-code: no web tool in the bundle, and the gate blocks a web call (1.607276ms) ✔ a granted session's gate lets the typed web tools through and still blocks the built-ins (1.491851ms) ✔ search-source searxng-url: a granted role gets the typed broker tools and no harness web tool; an ungranted role gets nothing (3.994382ms) ✔ search-source searxng-bundled: a granted role gets the typed broker tools and no harness web tool; an ungranted role gets nothing (4.607494ms) ✔ search-source harness-builtin: claude-code gets WebSearch and WebFetch and no typed web tool, only when granted (7.10498ms) ✔ search-source harness-builtin is refused for a granted pi role, which has no built-in web tool (1.963859ms) ✔ allow exits 0, a deny exits 2 with the reason on stderr (136.608806ms) ✔ a missing or wrong policy, or a bad event, exits 2 (80.428616ms) ✔ the bundle's wrapped command: a missing gate or node still blocks (1083.272246ms) ✔ claude: typed tools through MCP, the hook blocks, builtins outside --tools don't exist (809.201996ms) ✔ claude: the hook alone blocks a path outside the workspace (492.968668ms) ✔ claude: a second turn resumes the first turn's session (743.320282ms) ✔ claude adapter: --restricted is always passed (5.322758ms) ✔ claude adapter: MOSAIC_CLAUDE_BIN wins over PATH and must be an absolute executable (16.674144ms) ✔ claude adapter: a directory pin exits 2 before it runs; a pin with a space runs (5.584381ms) ✔ claude: CLAUDE.md files and auto-memory don't reach the model; without --restricted they do (731.222879ms) ✔ claude: a missing hook or MCP file refuses before claude starts (7.77514ms) ✔ pi: policy tools and typed tools pass, anything else is blocked (4.031057ms) ✔ claude: builtins map from pi names, typed tools need the mcp prefix (0.893263ms) ✔ file tool paths must resolve inside the workspace (1.51737ms) ✔ pi's own path normalisation can't be used to step out (1.134759ms) ✔ a symlink inside the workspace that points out is outside (1.230296ms) ✔ a dangling symlink is refused at any depth, in both harnesses (5.347227ms) ✔ read is checked under every spelling pi's read would open, in both harnesses (20.029458ms) ✔ other spellings cover directories, dangling links and pi's cwd (2.006651ms) ✔ a relative path climbs from the workspace's real path, in both harnesses (6.380537ms) ✔ claude path fields per tool (1.778925ms) ✔ glob patterns stay inside the workspace (1.88594ms) ✔ a path that can't be checked is blocked (0.458795ms) ✔ fetch_review's have lists the working refs before origin's, and no more than HAVE_MAX (235.925094ms) ✔ fetch_review's have orders the working refs newest first within each clone, one clone after another (147.389035ms) ✔ a runner git past its time: SIGTERM reaches its process group, and SIGKILL follows after the grace period (1808.811567ms) ✔ once a signalled runner git has closed, its group gets SIGKILL, so a process that ignored SIGTERM dies too (302.6597ms) ✔ land_work sends one land line and resolves the host's started answer (3.224586ms) ✔ initialize, ping and tools/list (58.635604ms) ✔ tools/call goes through the tool socket; a refusal is an isError result (48.439466ms) ✔ replies are compact JSON; over the cap a tool error names the size and the cap (39.641118ms) ✔ unknown tools and methods are JSON-RPC errors and never reach the socket (31.19734ms) ✔ a missing argument is a usage error (24.706001ms) ✔ pi: typed tools reach the socket, the gate blocks, agent_end writes the marker (389.253501ms) ✔ pi: a reply over the cap is a tool error naming the size and the cap; list_tasks asks for open tasks (330.067992ms) ✔ pi: a write through a dangling symlink is blocked, and nothing appears outside (319.172745ms) ✔ pi: a read is refused when pi would open another spelling outside (302.313433ms) ✔ pi: a relative path climbs from the real path of a workspace behind a symlink (298.076739ms) ✔ pi: a relative path climbs from the real path of a dataRoot behind a symlink (305.810518ms) ✔ pi: a missing extension refuses before any model call (6.18366ms) ✔ pi: an extension without its configuration fails pi's start (256.935667ms) ✔ founderCheck: founder variables, then a needed service without a usable token (1.626178ms) ✔ turnRequest names the sender, class, reply and decision (0.298951ms) ✔ turnRequest says a tester's message is a test and never Jason's instruction (0.301766ms) ✔ turnRequest shows the message's own sent time, a day old as well as new (0.2ms) ✔ stateBlock names the launch state with its time and event, and the open tasks with the cap (0.349308ms) ✔ a message becomes a turn, the answer goes back as a RESULT, SIGTERM releases and exits 0 (280.758678ms) ✔ launches revoked and then restored: the turn's block reads allowed and gives the restore time (230.484175ms) ✔ a task closed on the tracker is missing from the next turn's open list (230.839356ms) ✔ a message read a day after it was sent shows the time it was sent (184.473324ms) ✔ a decision the human resolves becomes a turn citing the decision and the answer (224.814103ms) ✔ a failed state read fails the turn: no adapter runs, the sender hears why, the message is read (119.528307ms) ✔ a SIGTERM before the claim stops the runner with exit 0 and no claim (105.392451ms) ✔ typed tools carry the runner's capability; launch goes to the host's launch socket (430.052564ms) ✔ end_run and change_organization reach the broker with the tool's arguments, and refusals keep their codes (429.257417ms) ✔ a RESULT gets no automatic reply; failed turns reply with the reason (1282.709813ms) ✔ the bus's launch notice is one turn with no reply, and is not taken again (251.272061ms) ✔ SIGTERM during a turn kills the turn's process group and still exits 0 (148.847029ms) ✔ founder credentials stop before the claim (20) (170.30994ms) ✔ a refused claim exits 21; an ended run's capability exits 22 (195.756258ms) ✔ the launch ending under a running session exits 22 (145.490512ms) ✔ a broker that stays unreachable exits 23 after brokerRetries polls (252.902448ms) ✔ a broker that is down at the claim exits 23, not 21 (88.491142ms) ✔ no capability, or a malformed one, on stdin exits 2 (191.406913ms) ✔ a missing or malformed policy exits 2 before the claim (122.254862ms) ✔ every role sees read_spec, record writer follows its grant, citation schemas stay strict (17.592288ms) ✔ typed socket forwards read content/history and agent record writes, and cannot write a spec (195.670224ms) ✔ the PM gets its task verbs and the reads, and no launch (10.560908ms) ✔ the chief of staff gets launch, send_message, end_run and change_organization, and no task verbs (4.040919ms) ✔ request_staff is the pm's alone: staff.request over message.send, roles and counts bounded (13.654609ms) ✔ end_run sends role and reason, no decision; change_organization sends revision and the organization keys (2.398039ms) ✔ the organization keys change_organization takes are the broker's ORGANIZATION_KEYS (0.171689ms) ✔ a role without role.end or organization.change gets neither tool, and each tool follows its own grant (9.199641ms) ✔ the assistant gets send_message, the reads and the web tools (it holds the web grant), no launch (1.093508ms) ✔ list_requirements reads the requirement ids, for a task.create holder with a tracker (2.138514ms) ✔ a coder gets no launch, no resolve_decision, and no task tools without a tracker (1.151429ms) ✔ a coder with a tracker gets its own task tools and no task_reassign (0.957224ms) ✔ launch only when the business's launch block names the instance as launcher (1.704609ms) ✔ an action outside the instance's authority has no tool (1.454172ms) ✔ the review loop: push_working and request_review for the coder, fetch_review and review_verdict for the reviewer, none for the PM (4.824057ms) ✔ land_work is there only with land, after push_working, with the task and the decision (3.115143ms) ✔ callTool: one JSON line out, the result back, a refusal rejects (8.947507ms) ✔ list_tasks takes the broker's filters and pages; get_task takes a task_ref (2.417558ms) ✔ callTool sends list_tasks done false unless the session names done; other tools as given (2.604895ms) ✔ replyText: compact JSON up to the cap, an error naming the size and the cap over it (0.604995ms) ✔ web_search's query maxLength is the web layer's SEARCH_MAX_QUERY_CHARS (2.565861ms) ℹ tests 108 ℹ suites 0 ℹ pass 108 ℹ fail 0 ℹ cancelled 0 ℹ skipped 0 ℹ todo 0 ℹ duration_ms 12062.67946