✔ sessionModel: agent vars win, then the system's execution settings (11.429864ms)
✔ a pi bundle: prompt, policy, tools and manifest, 0600 in a 0700 directory (6.067533ms)
✔ a chief-of-staff bundle lists launch among the bus tools and the manifest's typed tools (3.48531ms)
✔ a chief-of-staff bundle lists end_run and change_organization as typed and only instance.name untyped, for either harness (5.296467ms)
✔ the stop line names end_run for a role.end holder, and every other role keeps today's text (6.312914ms)
✔ a claude-code bundle adds the wrapped gate hook and the MCP config (1.929151ms)
✔ the prompt's built-in tools line is the harness's own names, as the manifest lists them (7.859121ms)
✔ the prompt names the project clones in the workspace, and none when there are none (6.683098ms)
✔ the prompt explains the review loop to a role with push_working, and reviews to one with fetch_review (7.867317ms)
✔ the prompt explains landing to a role with land_work, and to no other (5.417925ms)
✔ the prompt tells the pm to ask the chief of staff for sessions, and no other role (9.381567ms)
✔ a bundle is written once: an existing file refuses (2.260321ms)
✔ a path with a single quote can't go into the hook command (2.24652ms)
✔ web grant on pi: the bundle has web_search and web_fetch, and no built-in web tool (2.799335ms)
✔ no web grant on pi: no web tool in the bundle, and the gate blocks a web call (2.281637ms)
✔ web grant on claude-code: the bundle has web_search and web_fetch, and no built-in web tool (1.916494ms)
✔ no web grant on claude-code: no web tool in the bundle, and the gate blocks a web call (2.201772ms)
✔ a granted session's gate lets the typed web tools through and still blocks the built-ins (2.400824ms)
✔ search-source searxng-url: a granted role gets the typed broker tools and no harness web tool; an ungranted role gets nothing (4.611736ms)
✔ search-source searxng-bundled: a granted role gets the typed broker tools and no harness web tool; an ungranted role gets nothing (5.43772ms)
✔ search-source harness-builtin: claude-code gets WebSearch and WebFetch and no typed web tool, only when granted (6.484007ms)
✔ search-source harness-builtin is refused for a granted pi role, which has no built-in web tool (1.734069ms)
✔ allow exits 0, a deny exits 2 with the reason on stderr (137.980447ms)
✔ a missing or wrong policy, or a bad event, exits 2 (90.618991ms)
✔ the bundle's wrapped command: a missing gate or node still blocks (1082.825292ms)
✔ claude: typed tools through MCP, the hook blocks, builtins outside --tools don't exist (761.375671ms)
✔ claude: the hook alone blocks a path outside the workspace (404.38052ms)
✔ claude: a second turn resumes the first turn's session (699.377475ms)
✔ claude adapter: --restricted is always passed (4.860152ms)
✔ claude adapter: MOSAIC_CLAUDE_BIN wins over PATH and must be an absolute executable (22.830569ms)
✔ claude adapter: a directory pin exits 2 before it runs; a pin with a space runs (5.643804ms)
✔ claude: CLAUDE.md files and auto-memory don't reach the model; without --restricted they do (681.050168ms)
✔ claude: a missing hook or MCP file refuses before claude starts (7.45554ms)
✔ pi: policy tools and typed tools pass, anything else is blocked (4.352273ms)
✔ claude: builtins map from pi names, typed tools need the mcp prefix (0.72111ms)
✔ file tool paths must resolve inside the workspace (1.014185ms)
✔ pi's own path normalisation can't be used to step out (1.03841ms)
✔ a symlink inside the workspace that points out is outside (0.845267ms)
✔ a dangling symlink is refused at any depth, in both harnesses (3.235167ms)
✔ read is checked under every spelling pi's read would open, in both harnesses (19.716419ms)
✔ other spellings cover directories, dangling links and pi's cwd (2.106785ms)
✔ a relative path climbs from the workspace's real path, in both harnesses (6.727631ms)
✔ claude path fields per tool (1.029206ms)
✔ glob patterns stay inside the workspace (0.781392ms)
✔ a path that can't be checked is blocked (0.5137ms)
✔ fetch_review's have lists the working refs before origin's, and no more than HAVE_MAX (246.38436ms)
✔ fetch_review's have orders the working refs newest first within each clone, one clone after another (123.52209ms)
✔ a runner git past its time: SIGTERM reaches its process group, and SIGKILL follows after the grace period (1809.782337ms)
✔ once a signalled runner git has closed, its group gets SIGKILL, so a process that ignored SIGTERM dies too (303.776664ms)
✔ land_work sends one land line and resolves the host's started answer (3.161815ms)
✔ initialize, ping and tools/list (45.665575ms)
✔ tools/call goes through the tool socket; a refusal is an isError result (42.027664ms)
✔ replies are compact JSON; over the cap a tool error names the size and the cap (44.583773ms)
✔ unknown tools and methods are JSON-RPC errors and never reach the socket (37.775098ms)
✔ a missing argument is a usage error (27.78227ms)
✔ pi: typed tools reach the socket, the gate blocks, agent_end writes the marker (389.471191ms)
✔ pi: a reply over the cap is a tool error naming the size and the cap; list_tasks asks for open tasks (321.187531ms)
✔ pi: a write through a dangling symlink is blocked, and nothing appears outside (317.610462ms)
✔ pi: a read is refused when pi would open another spelling outside (307.078952ms)
✔ pi: a relative path climbs from the real path of a workspace behind a symlink (327.110136ms)
✔ pi: a relative path climbs from the real path of a dataRoot behind a symlink (295.276354ms)
✔ pi: a missing extension refuses before any model call (6.080422ms)
✔ pi: an extension without its configuration fails pi's start (246.386515ms)
✔ founderCheck: founder variables, then a needed service without a usable token (1.135264ms)
✔ turnRequest names the sender, class, reply and decision (0.21951ms)
✔ turnRequest says a tester's message is a test and never Jason's instruction (0.232866ms)
✔ turnRequest shows the message's own sent time, a day old as well as new (0.156017ms)
✔ stateBlock names the launch state with its time and event, and the open tasks with the cap (0.263507ms)
✔ a message becomes a turn, the answer goes back as a RESULT, SIGTERM releases and exits 0 (191.026636ms)
✔ launches revoked and then restored: the turn's block reads allowed and gives the restore time (178.224872ms)
✔ a task closed on the tracker is missing from the next turn's open list (196.59284ms)
✔ a message read a day after it was sent shows the time it was sent (90.828023ms)
✔ a decision the human resolves becomes a turn citing the decision and the answer (144.023778ms)
✔ a failed state read fails the turn: no adapter runs, the sender hears why, the message is read (66.07603ms)
✔ a SIGTERM before the claim stops the runner with exit 0 and no claim (51.683392ms)
✔ typed tools carry the runner's capability; launch goes to the host's launch socket (276.479253ms)
✔ end_run and change_organization reach the broker with the tool's arguments, and refusals keep their codes (259.292191ms)
✔ a RESULT gets no automatic reply; failed turns reply with the reason (1198.34556ms)
✔ the bus's launch notice is one turn with no reply, and is not taken again (215.872019ms)
✔ SIGTERM during a turn kills the turn's process group and still exits 0 (93.571183ms)
✔ founder credentials stop before the claim (20) (89.693727ms)
✔ a refused claim exits 21; an ended run's capability exits 22 (96.347337ms)
✔ the launch ending under a running session exits 22 (105.656795ms)
✔ a broker that stays unreachable exits 23 after brokerRetries polls (201.681021ms)
✔ a broker that is down at the claim exits 23, not 21 (45.15795ms)
✔ no capability, or a malformed one, on stdin exits 2 (148.414599ms)
✔ a missing or malformed policy exits 2 before the claim (82.49756ms)
✔ every role sees read_spec, record writer follows its grant, citation schemas stay strict (13.658463ms)
✔ typed socket forwards read content/history and agent record writes, and cannot write a spec (102.130833ms)
✔ the PM gets its task verbs and the reads, and no launch (9.482525ms)
✔ send_message offers the BLOCK class, beside the classes it already had (3.33285ms)
✔ the chief of staff gets launch, send_message, end_run and change_organization, and no task verbs (2.343953ms)
✔ request_staff is the pm's alone: staff.request over message.send, roles and counts bounded (11.025538ms)
✔ end_run sends role and reason, no decision; change_organization sends revision and the organization keys (1.42881ms)
✔ the organization keys change_organization takes are the broker's ORGANIZATION_KEYS (0.098128ms)
✔ a role without role.end or organization.change gets neither tool, and each tool follows its own grant (7.029505ms)
✔ the assistant gets send_message, the reads and the web tools (it holds the web grant), no launch (1.186894ms)
✔ list_requirements reads the requirement ids, for a task.create holder with a tracker (2.482733ms)
✔ a coder gets no launch, no resolve_decision, and no task tools without a tracker (1.149525ms)
✔ a coder with a tracker gets its own task tools and no task_reassign (0.993306ms)
✔ launch only when the business's launch block names the instance as launcher (2.158823ms)
✔ an action outside the instance's authority has no tool (1.302845ms)
✔ the review loop: push_working and request_review for the coder, fetch_review and review_verdict for the reviewer, none for the PM (4.526727ms)
✔ land_work is there only with land, after push_working, with the task and the decision (3.057475ms)
✔ callTool: one JSON line out, the result back, a refusal rejects (13.60265ms)
✔ list_tasks takes the broker's filters and pages; get_task takes a task_ref (3.345792ms)
✔ callTool sends list_tasks done false unless the session names done; other tools as given (3.53927ms)
✔ replyText: compact JSON up to the cap, an error naming the size and the cap over it (0.95941ms)
✔ web_search's query maxLength is the web layer's SEARCH_MAX_QUERY_CHARS (3.518638ms)
ℹ tests 109
ℹ suites 0
ℹ pass 109
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 11095.250354
