✔ sessionModel: agent vars win, then the system's execution settings (11.932961ms)
✔ a pi bundle: prompt, policy, tools and manifest, 0600 in a 0700 directory (4.66256ms)
✔ a chief-of-staff bundle lists launch among the bus tools and the manifest's typed tools (2.358504ms)
✔ a chief-of-staff bundle lists end_run and change_organization as typed and only instance.name untyped, for either harness (4.957887ms)
✔ the stop line names end_run for a role.end holder, and every other role keeps today's text (6.967488ms)
✔ a claude-code bundle adds the wrapped gate hook and the MCP config (3.120911ms)
✔ the prompt's built-in tools line is the harness's own names, as the manifest lists them (7.829983ms)
✔ the prompt names the project clones in the workspace, and none when there are none (8.80054ms)
✔ the prompt explains the review loop to a role with push_working, and reviews to one with fetch_review (8.052814ms)
✔ the prompt explains landing to a role with land_work, and to no other (4.367229ms)
✔ the prompt tells the pm to ask the chief of staff for sessions, and no other role (8.112295ms)
✔ a bundle is written once: an existing file refuses (1.613938ms)
✔ a path with a single quote can't go into the hook command (1.593316ms)
✔ web grant on pi: the bundle has web_search and web_fetch, and no built-in web tool (1.762196ms)
✔ no web grant on pi: no web tool in the bundle, and the gate blocks a web call (1.727278ms)
✔ web grant on claude-code: the bundle has web_search and web_fetch, and no built-in web tool (1.666807ms)
✔ no web grant on claude-code: no web tool in the bundle, and the gate blocks a web call (1.711376ms)
✔ a granted session's gate lets the typed web tools through and still blocks the built-ins (1.476171ms)
✔ search-source searxng-url: a granted role gets the typed broker tools and no harness web tool; an ungranted role gets nothing (3.892635ms)
✔ search-source searxng-bundled: a granted role gets the typed broker tools and no harness web tool; an ungranted role gets nothing (4.367349ms)
✔ search-source harness-builtin: claude-code gets WebSearch and WebFetch and no typed web tool, only when granted (6.005822ms)
✔ search-source harness-builtin is refused for a granted pi role, which has no built-in web tool (1.603989ms)
✔ allow exits 0, a deny exits 2 with the reason on stderr (131.7789ms)
✔ a missing or wrong policy, or a bad event, exits 2 (81.446065ms)
✔ the bundle's wrapped command: a missing gate or node still blocks (1089.369438ms)
✔ claude: typed tools through MCP, the hook blocks, builtins outside --tools don't exist (738.140526ms)
✔ claude: the hook alone blocks a path outside the workspace (405.18525ms)
✔ claude: a second turn resumes the first turn's session (693.593913ms)
✔ claude adapter: --restricted is always passed (5.387426ms)
✔ claude adapter: MOSAIC_CLAUDE_BIN wins over PATH and must be an absolute executable (19.459102ms)
✔ claude adapter: a directory pin exits 2 before it runs; a pin with a space runs (5.851978ms)
✔ claude: CLAUDE.md files and auto-memory don't reach the model; without --restricted they do (689.786048ms)
✔ claude: a missing hook or MCP file refuses before claude starts (7.097216ms)
✔ pi: policy tools and typed tools pass, anything else is blocked (2.932331ms)
✔ claude: builtins map from pi names, typed tools need the mcp prefix (0.713859ms)
✔ file tool paths must resolve inside the workspace (1.063612ms)
✔ pi's own path normalisation can't be used to step out (0.885872ms)
✔ a symlink inside the workspace that points out is outside (0.762992ms)
✔ a dangling symlink is refused at any depth, in both harnesses (2.794471ms)
✔ read is checked under every spelling pi's read would open, in both harnesses (15.797177ms)
✔ other spellings cover directories, dangling links and pi's cwd (1.75855ms)
✔ a relative path climbs from the workspace's real path, in both harnesses (5.22344ms)
✔ claude path fields per tool (0.693391ms)
✔ glob patterns stay inside the workspace (0.639546ms)
✔ a path that can't be checked is blocked (0.387799ms)
✔ fetch_review's have lists the working refs before origin's, and no more than HAVE_MAX (227.650297ms)
✔ fetch_review's have orders the working refs newest first within each clone, one clone after another (136.331206ms)
✔ a runner git past its time: SIGTERM reaches its process group, and SIGKILL follows after the grace period (1810.092704ms)
✔ once a signalled runner git has closed, its group gets SIGKILL, so a process that ignored SIGTERM dies too (304.108566ms)
✔ land_work sends one land line and resolves the host's started answer (3.293669ms)
✔ initialize, ping and tools/list (47.824063ms)
✔ tools/call goes through the tool socket; a refusal is an isError result (49.490599ms)
✔ replies are compact JSON; over the cap a tool error names the size and the cap (38.795634ms)
✔ unknown tools and methods are JSON-RPC errors and never reach the socket (31.164372ms)
✔ a missing argument is a usage error (29.429372ms)
✔ pi: typed tools reach the socket, the gate blocks, agent_end writes the marker (379.121246ms)
✔ pi: a reply over the cap is a tool error naming the size and the cap; list_tasks asks for open tasks (307.113718ms)
✔ pi: a write through a dangling symlink is blocked, and nothing appears outside (314.876716ms)
✔ pi: a read is refused when pi would open another spelling outside (303.642912ms)
✔ pi: a relative path climbs from the real path of a workspace behind a symlink (301.72971ms)
✔ pi: a relative path climbs from the real path of a dataRoot behind a symlink (304.958343ms)
✔ pi: a missing extension refuses before any model call (6.235441ms)
✔ pi: an extension without its configuration fails pi's start (251.868227ms)
✔ founderCheck: founder variables, then a needed service without a usable token (1.753668ms)
✔ turnRequest names the sender, class, reply and decision (0.295609ms)
✔ turnRequest says a tester's message is a test and never Jason's instruction (0.295372ms)
✔ turnRequest shows the message's own sent time, a day old as well as new (0.200462ms)
✔ stateBlock names the launch state with its time and event, and the open tasks with the cap (0.326406ms)
✔ a message becomes a turn, the answer goes back as a RESULT, SIGTERM releases and exits 0 (142.383827ms)
✔ launches revoked and then restored: the turn's block reads allowed and gives the restore time (148.619594ms)
✔ a task closed on the tracker is missing from the next turn's open list (175.711949ms)
✔ a message read a day after it was sent shows the time it was sent (90.399663ms)
✔ a decision the human resolves becomes a turn citing the decision and the answer (143.009058ms)
✔ a failed state read fails the turn: no adapter runs, the sender hears why, the message is read (63.312882ms)
✔ a SIGTERM before the claim stops the runner with exit 0 and no claim (54.643174ms)
✔ typed tools carry the runner's capability; launch goes to the host's launch socket (279.240728ms)
✔ end_run and change_organization reach the broker with the tool's arguments, and refusals keep their codes (258.329848ms)
✔ a RESULT gets no automatic reply; failed turns reply with the reason (1172.811502ms)
✔ the bus's launch notice is one turn with no reply, and is not taken again (217.806873ms)
✔ SIGTERM during a turn kills the turn's process group and still exits 0 (95.298643ms)
✔ founder credentials stop before the claim (20) (92.714683ms)
✔ a refused claim exits 21; an ended run's capability exits 22 (97.988882ms)
✔ the launch ending under a running session exits 22 (101.240597ms)
✔ a broker that stays unreachable exits 23 after brokerRetries polls (201.569556ms)
✔ a broker that is down at the claim exits 23, not 21 (58.659251ms)
✔ no capability, or a malformed one, on stdin exits 2 (151.225496ms)
✔ a missing or malformed policy exits 2 before the claim (91.625182ms)
✔ every role sees read_spec, record writer follows its grant, citation schemas stay strict (14.347546ms)
✔ typed socket forwards read content/history and agent record writes, and cannot write a spec (95.471128ms)
✔ the PM gets its task verbs and the reads, and no launch (8.104158ms)
✔ send_message offers the BLOCK class, beside the classes it already had (3.087465ms)
✔ the chief of staff gets launch, send_message, end_run and change_organization, and no task verbs (1.791094ms)
✔ request_staff is the pm's alone: staff.request over message.send, roles and counts bounded (10.330793ms)
✔ end_run sends role and reason, no decision; change_organization sends revision and the organization keys (1.542871ms)
✔ the organization keys change_organization takes are the broker's ORGANIZATION_KEYS (0.092683ms)
✔ a role without role.end or organization.change gets neither tool, and each tool follows its own grant (6.553957ms)
✔ the assistant gets send_message, the reads and the web tools (it holds the web grant), no launch (1.200829ms)
✔ list_requirements reads the requirement ids, for a task.create holder with a tracker (2.335249ms)
✔ a coder gets no launch, no resolve_decision, and no task tools without a tracker (1.227262ms)
✔ a coder with a tracker gets its own task tools and no task_reassign (1.14268ms)
✔ launch only when the business's launch block names the instance as launcher (1.896477ms)
✔ an action outside the instance's authority has no tool (1.246415ms)
✔ the review loop: push_working and request_review for the coder, fetch_review and review_verdict for the reviewer, none for the PM (4.708034ms)
✔ land_work is there only with land, after push_working, with the task and the decision (2.385657ms)
✔ callTool: one JSON line out, the result back, a refusal rejects (9.859959ms)
✔ list_tasks takes the broker's filters and pages; get_task takes a task_ref (2.368822ms)
✔ callTool sends list_tasks done false unless the session names done; other tools as given (2.5048ms)
✔ replyText: compact JSON up to the cap, an error naming the size and the cap over it (0.595655ms)
✔ web_search's query maxLength is the web layer's SEARCH_MAX_QUERY_CHARS (2.495151ms)
ℹ tests 109
ℹ suites 0
ℹ pass 109
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 10966.251242
