✔ sessionModel: agent vars win, then the system's execution settings (15.097602ms)
✔ a pi bundle: prompt, policy, tools and manifest, 0600 in a 0700 directory (6.097868ms)
✔ a claude-code bundle adds the wrapped gate hook and the MCP config (3.358165ms)
✔ a bundle is written once: an existing file refuses (2.910743ms)
✔ a path with a single quote can't go into the hook command (2.716976ms)
✔ allow exits 0, a deny exits 2 with the reason on stderr (198.562157ms)
✔ a missing or wrong policy, or a bad event, exits 2 (125.951469ms)
✔ the bundle's wrapped command: a missing gate or node still blocks (1138.55797ms)
✔ claude: typed tools through MCP, the hook blocks, builtins outside --tools don't exist (1449.8132ms)
✔ claude: the hook alone blocks a path outside the workspace (552.581946ms)
✔ claude: a second turn resumes the first turn's session (887.521272ms)
✔ claude adapter: --restricted is always passed (7.845157ms)
✔ claude adapter: MOSAIC_CLAUDE_BIN wins over PATH and must be an absolute executable (36.152914ms)
✔ claude adapter: a directory pin exits 2 before it runs; a pin with a space runs (11.658979ms)
✔ claude: CLAUDE.md files and auto-memory don't reach the model; without --restricted they do (945.726312ms)
✔ claude: a missing hook or MCP file refuses before claude starts (10.879578ms)
✔ pi: policy tools and typed tools pass, anything else is blocked (3.997626ms)
✔ claude: builtins map from pi names, typed tools need the mcp prefix (1.095666ms)
✔ file tool paths must resolve inside the workspace (1.903989ms)
✔ pi's own path normalisation can't be used to step out (1.523161ms)
✔ a symlink inside the workspace that points out is outside (0.986938ms)
✔ a dangling symlink is refused at any depth, in both harnesses (4.595008ms)
✔ read is checked under every spelling pi's read would open, in both harnesses (24.131388ms)
✔ other spellings cover directories, dangling links and pi's cwd (1.889138ms)
✔ a relative path climbs from the workspace's real path, in both harnesses (7.448035ms)
✔ claude path fields per tool (2.25162ms)
✔ glob patterns stay inside the workspace (0.810033ms)
✔ a path that can't be checked is blocked (0.375567ms)
✔ initialize, ping and tools/list (66.981066ms)
✔ tools/call goes through the tool socket; a refusal is an isError result (54.207372ms)
✔ unknown tools and methods are JSON-RPC errors and never reach the socket (52.929411ms)
✔ a missing argument is a usage error (53.671989ms)
✔ pi: typed tools reach the socket, the gate blocks, agent_end writes the marker (573.847505ms)
✔ pi: a write through a dangling symlink is blocked, and nothing appears outside (397.472721ms)
✔ pi: a read is refused when pi would open another spelling outside (343.928275ms)
✔ pi: a relative path climbs from the real path of a workspace behind a symlink (393.322868ms)
✔ pi: a relative path climbs from the real path of a dataRoot behind a symlink (417.483266ms)
✔ pi: a missing extension refuses before any model call (8.185612ms)
✔ pi: an extension without its configuration fails pi's start (335.304274ms)
✔ founderCheck: founder variables, then a needed service without a usable token (2.517134ms)
✔ turnRequest names the sender, class, reply and decision (0.333266ms)
✔ a message becomes a turn, the answer goes back as a RESULT, SIGTERM releases and exits 0 (228.464532ms)
✔ a SIGTERM before the claim stops the runner with exit 0 and no claim (81.232047ms)
✔ typed tools carry the runner's capability; launch goes to the host's launch socket (390.880481ms)
✔ a RESULT gets no automatic reply; failed turns reply with the reason (1224.709889ms)
✔ SIGTERM during a turn kills the turn's process group and still exits 0 (124.494814ms)
✔ founder credentials stop before the claim (20) (120.95178ms)
✔ a refused claim exits 21; an ended run's capability exits 22 (125.318769ms)
✔ the launch ending under a running session exits 22 (115.856272ms)
✔ a broker that stays unreachable exits 23 after brokerRetries polls (213.292434ms)
✔ a broker that is down at the claim exits 23, not 21 (63.560481ms)
✔ no capability, or a malformed one, on stdin exits 2 (202.236177ms)
✔ a missing or malformed policy exits 2 before the claim (128.552785ms)
✔ the PM gets launch, its task verbs and the reads (10.750162ms)
✔ a coder gets no launch, no resolve_decision, and no task tools without a tracker (3.109782ms)
✔ a coder with a tracker gets its own task tools and no task_reassign (2.453379ms)
✔ launch only when the business's launch block names the instance as launcher (3.203499ms)
✔ an action outside the instance's authority has no tool (3.229567ms)
✔ callTool: one JSON line out, the result back, a refusal rejects (12.162075ms)
ℹ tests 59
ℹ suites 0
ℹ pass 59
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 10229.260926