pi-crew 0.9.62 → 0.9.64

package/CHANGELOG.md

@@ -2,6 +2,98 @@
 
 > **Note:** `atomic-write-v2.ts` / `AtomicWriter` mentioned in historical entries below was consolidated into `atomic-write.ts` as of v0.9.42. This changelog is preserved as historical record — the migration was completed (the v2 class was never adopted; v1 won on simplicity + symlink-safety + link+unlink atomicity). See `docs/migration/atomic-write-v2-migration.md` for the decision rationale.
 
+## [0.9.64] — pi-rlm→pi-crew pattern transfer: worker scratchpad + crash-resume + cancellation + quick wins (2026-08-09)
+
+### Quick Wins (patterns 17/19/20/11 + spike CI)
+
+### Features
+- **Schema-driven docs (QW17)**: `agents/*.md` frontmatter `tools:` now matches the
+  enforced `ROLE_TOOL_CONFIGS` (4 drifts fixed); a sync-test pins the derivation;
+  `scripts/gen-role-tools-docs.mjs` renders `docs/role-tools.md` from the source.
+- **Retry-resume contract suite (QW19)**: pins `executeWithRetry` (`?`-glob, empty-
+  retryableErrors, maxAttempts:0, abort-during-sleep, default attemptId),
+  `FileCheckpointStore` (corrupt-file quarantine, list-skip, wrong-runId delete),
+  and exports `shouldRecoverTask` for direct testing.
+- **Failure-mode inventory (QW20)**: `docs/failure-mode-inventory.md` maps the 7
+  pi-rlm failure modes to pi-crew handlers (wedge gap closed by Phase 1
+  ping-before-execute; EPIPE/timeout interplay declared as gaps).
+- **Error-as-data contract (QW11)**: `evidenceStatusFor` + `attemptErrorFor`
+  extracted as pure functions from `runChildProcessTask`; contract test pins the
+  precedence (cancelled > failed > completed; E007 timedOut override; 429 gated).
+- **`test:spike` script**: wires the scratchpad spike tests into CI.
+
+
+### Phase 3 cancellation hardening (experimental)
+
+### Features
+
+- **Atomic snapshot write (D2')**: `engine.snapshotState` now writes via
+  `temp + rename` (same directory), eliminating the theoretical torn-write race
+  between the debounce timer and the F3 quit flush.
+- **Kill-and-restore verified**: the existing SIGTERM → pi print-mode dispose →
+  `session_shutdown` reason:"quit" → F3 flush chain already captures a fresh
+  snapshot on worker kill (no new handler required). Documented + pinned by a
+  gated real-worker test.
+- **EngineBusyError: deliberately skipped** — ping-before-execute already
+  detects a wedged guest; the FIFO execute queue makes a busy-reject redundant.
+
+
+### Phase 2 crash-resume (experimental)
+
+### Features
+
+- **Cross-attempt restore (crash-resume)**: a scratchpad worker attempt N+1 (retry
+  / crash-recovery re-queue / manual re-run) automatically revives the namespace
+  from the previous attempt's redacted snapshot artifact — via a single
+  spawn-time lookup (`findLatestScratchpadSnapshot`), no retry-loop / recovery
+  changes needed. Restores lazily on the first `execute` call (D7 lazy invariant
+  kept) with a one-line model-visible notice.
+  - Lookup: latest mtime (model-fallback index resets each retry round → number
+    is not write-order); tie-break lowest attempt = newest round.
+  - Security: read-time re-validation (containment + filename pattern + lstat +
+    size + mtime hint, D10); fail-open (D11) never breaks the worker; strict scan
+    (D12); base64 round-trip (D13); redacted secret → literal `"***"` (D4).
+- **Guest zombie backstop (SEC-1)**: the engine now overrides
+  `PI_CREW_PARENT_PID=worker pid` for the guest (D5), so an orphaned guest
+  (worker SIGKILL'd) is flagged by the zombie scanner — pure inheritance left
+  guests LIVE forever holding provider keys + broker token.
+- **Snapshot cap (D6)**: 4 MiB two-sided (write-side raw byteLength trim, read-
+  side file + per-var 256 KiB) bounds v8.deserialize amplification.
+
+
+### Phase 1 worker scratchpad (experimental)
+
+### Features
+
+- **Worker stateful scratchpad (experimental)**: opt-in persistent Bun-free JS
+  evaluator (`execute` tool) for workers, ported from the pi-rlm pattern.
+  State (variables, parsed data) compounds across `execute` calls within a task
+  attempt — intermediate results live in a namespace instead of being re-derived
+  from transcript text. Snapshot per-attempt into the run artifact store
+  (redacted + atomic) prepares the ground for Phase 2 crash-resume.
+  - Opt-in per role: `executor`, `test-engineer`, `verifier` (default on); other
+    roles via agent frontmatter `scratchpad: true` (write roles only).
+  - Security: S-6 read-only roles are gated out regardless of frontmatter
+    (privilege-elevation guard); F6 `scratchpad: false` is an explicit kill-switch;
+    raw snapshots never land in the artifact root (temp dir → redacted
+    `writeArtifact`).
+  - Dormant by default: only active when `PI_CREW_SCRATCHPAD=1` is set by the
+    spawner; zero behavior change for non-opt-in workers.
+
+
+## [0.9.63] — built-in performance observability + local-path provider-extension discovery (2026-08-08)
+
+### Features
+
+- **Built-in performance observability (byte-built-in, always-on).** Every team run now auto-attaches a detached resource sampler and auto-generates a detailed performance report on completion — no separate benchmark harness needed. Toggle per-team via frontmatter `observability: true|false` (default `true` for parsed team files; direct-object `TeamConfig` fixtures stay unset for test isolation).
+  - **Live resource sampler** (`scripts/resource-sampler.mjs`): samples CPU/RSS per-PID every 2s via ppid-tree attribution (root runner + all child workers, including respawns), with PID-reuse guard (`/proc` starttime), first-sample CPU exclusion, and **6 live warning categories** — `high_cpu` (≥300%), `rss_jump` (+200MB/interval), `rss_high` (≥1GB), `rss_leak` (window-30 monotonic +100MB), `proc_died`, `proc_zombie`. Rate-limited (10s/pid/category); `--no-live-warn` flag to silence.
+  - **Post-hoc analyzer** (`scripts/analyze-run.mjs`): combines `events.jsonl` + transcripts + `resources.jsonl` into a markdown report at `docs/perf-report-<runId>.md` with **22 anomaly categories** (task_failed, model_retry/model_cascade, large_gap, slow_phase, launch_delay, drain_stall, token_imbalance, no_cache, worker_respawn_churn, api_error_storm, zero_output_completion, sustained_cpu, transient_cpu/rss_spike, rss_growth, run_not_completed, high_failure_rate, run_idle, cost_unreported, missing_transcript, sampler_gap, tool_churn), a per-subagent timeline (launch / respawn / startup / active-work / drain / finalize), and token/cost/model attribution. Optional `--agents` flag emits per-agent breakdowns.
+  - **Runtime wiring** (`src/runtime/team-runner.ts`): `startPerfSampler` spawns the sampler detached + `unref`'d (death never affects the run); `schedulePerfAnalyze` runs the analyzer +3s after `after_run_complete` via an `unref`'d `setTimeout` (never blocks run completion). Strict `observability !== true` keeps test fixtures from spawning. The sampler auto-stops when the run manifest reaches a terminal status (`--watch-run` mode).
+  - **Overhead ≈ 0** (measured A/B): sampler ~0.05% of one core, 56MB RSS fixed; analyzer ~72ms one-shot after run; ~32KB artifacts/run. Verified end-to-end on real runs — see `docs/real-test/reports/real-test-2026-08-07-perf-obs-overhead.md`.
+
+### Bug fixes
+
+- **Local-path provider extensions were not discovered for child workers (oc-go and any `pi install <local-path>` provider went "Model not found").** `discoverProviderExtensions` only resolved `npm:` specs from `~/.pi/agent/settings.json` `packages`, skipping local-path specs (e.g. `../../source/my_pi/source/pi-other-provider`) on the assumption that local paths were the pi-crew extension itself. That assumption was wrong for local provider extensions: `--no-extensions` in `buildPiWorkerArgs` stripped the provider from every child spawn → every model from that provider hit `Error: Model "…" not found` → the fallback chain burned 5 spawn-fails (~10s) per task before landing on a builtin provider. Fix: resolve local-path specs (`./`, `../`, absolute) relative to the settings.json dir — same trust level as `npm:` (settings packages are a sanctioned channel written by `pi install`); skip the pi-crew package itself via `packageRoot()` (a worker must not re-load the orchestrator). **SEC-1 preserved** (project/project-pi AGENT extensions in `.crew/agents/*.md` frontmatter stay gated — that is a separate, untrusted channel). Tests in `test/unit/runtime/model/provider-extensions.test.ts`. Investigation + correction of the earlier "hidden models" mis-attribution in `docs/real-test/reports/real-test-2026-08-08-provider-ext-local-path.md`.
 
 ## [0.9.62] — provider-quota attribution per live-session agent + dead-worker alert re-fire fix (2026-08-06)
 

package/README.md

@@ -69,6 +69,7 @@ repo: https://github.com/baphuongna/pi-crew
 - **Durable event replay** (L1, v0.9.8) — `RunEventBus.onWithReplay()` catches up a re-subscribing dashboard/overlay with events it missed during transient absence (toggle, reconnect), replaying from the durable JSONL log with seq-based dedup. No information loss even if the live subscriber was briefly gone.
 - **Lossless-by-default output handling** (L4, v0.9.8) — worker output thresholds sized from measured data (100% of real outputs fit without compaction); when compaction is unavoidable it keeps head+tail (preserves closing code fences/headings) instead of head-only truncation. No more `[pi-crew compacted N chars]` markers eating the end of a worker's result.
 - **Inter-pi broker** (v0.9.47, default-on) — a Unix-domain-socket message bus that lets concurrently-running Pi sessions pass messages, steering notes, and task-status events to each other. **On by default** on Linux + macOS; auto-disabled on native Windows (no unix socket). Three independent kill switches: `broker.enabled: false` (config), `PI_CREW_BROKER=0` (env, always wins), Windows auto-disable. See [docs/decisions/2026-07-22-broker-phase4-gated-on.md](docs/decisions/2026-07-22-broker-phase4-gated-on.md).
+- **Worker stateful scratchpad** (experimental, opt-in per role) — `executor` / `test-engineer` / `verifier` workers get a `scratchpad` tool: a persistent Bun-free JS evaluator whose namespace **compounds across calls within a task attempt** (variables set in one cell are visible in the next), so intermediate results live in memory instead of being re-derived from the transcript. Snapshots are flushed (redacted, atomic) per-attempt into the artifact store; the next attempt (retry / crash-recovery re-queue / re-run) **automatically revives the namespace** from the latest snapshot. Dormant by default (armed only when the spawner sets `PI_CREW_SCRATCHPAD=1`); zero behavior change for non-opt-in workers. Ported from the `@shift-labs/pi-rlm` pattern. See [src/runtime/scratchpad/README.md](src/runtime/scratchpad/README.md) (Phase 1-3 design, env keys, guards, threat model).
 - **`test:critical` + `real-test-pi-crew` skill** (v0.9.47) — a curated 14-file / 97-test subset (`npm run test:critical`, ~20s) for fast in-loop verification, plus a bundled skill distilling the full 8-tier end-to-end verification discipline (unit → 3-path kill-switch proof → typecheck/bundle → live TUI probing → smoke team run). Prevents the verifier-worker hang that full `npm test` (>4 min) caused against the 300s worker timeout.
 - **Provider extensions in subagents** (v0.9.57) — pi-crew spawns child-pi workers with `--no-extensions` (security posture), which made extension-registered providers (e.g. `pi-commandcode-provider`) unresolvable inside subagents. pi-crew now **auto-discovers provider packages** from `~/.pi/agent/settings.json` `packages` (npm: specs) and loads them via `--extension` in every builtin/user subagent — so **all provider models work in subagents**. An explicit `runtime.agentExtensions: string[]` config is an optional extra allowlist on top of auto-discovery. **SEC-1 preserved:** project/project-pi agents never receive these (env-gate unchanged).
 

package/agents/critic.md

@@ -5,7 +5,7 @@ model: false
 systemPromptMode: replace
 inheritProjectContext: true
 inheritSkills: false
-tools: read, grep, find, ls
+tools: read, grep, find, ls, glob
 ---
 
 You are a critical reviewer. Find flaws, missing steps, unsafe assumptions, overengineering, underengineering, and verification gaps. Return concrete fixes to the plan.

package/agents/explorer.md

@@ -5,7 +5,7 @@ model: false
 systemPromptMode: replace
 inheritProjectContext: true
 inheritSkills: false
-tools: read, grep, find, ls
+tools: read, grep, find, ls, glob, bash
 ---
 
 You are a fast codebase explorer. Map relevant files, symbols, data flow, and constraints. Do not modify files. Return concise findings with paths and evidence.

package/agents/planner.md

@@ -5,7 +5,7 @@ model: false
 systemPromptMode: replace
 inheritProjectContext: true
 inheritSkills: false
-tools: read, grep, find, ls
+tools: read, grep, find, ls, glob
 ---
 
 You are a planning specialist. Convert the goal and discovery notes into a concrete, ordered plan. Identify dependencies, risks, validation steps, and handoff instructions for implementers.

package/agents/reviewer.md

@@ -5,7 +5,7 @@ model: false
 systemPromptMode: replace
 inheritProjectContext: true
 inheritSkills: false
-tools: read, grep, find, ls, bash
+tools: read, grep, find, ls, glob, bash
 ---
 
 You are a code reviewer. Review the implementation for bugs, regressions, maintainability issues, missing tests, and project-rule violations. Return prioritized findings with evidence.

package/agents/security-reviewer.md

@@ -5,7 +5,7 @@ model: false
 systemPromptMode: replace
 inheritProjectContext: true
 inheritSkills: false
-tools: read, grep, find, ls, bash
+tools: read, grep, find
 ---
 
 You are a security reviewer. Look for injection, authn/authz flaws, insecure defaults, secret exposure, unsafe filesystem/network behavior, and dependency risks. Return severity and remediation.

package/agents/test-engineer.md

@@ -5,7 +5,7 @@ model: false
 systemPromptMode: replace
 inheritProjectContext: true
 inheritSkills: false
-tools: read, grep, find, ls, bash, edit, write
+tools: read, edit, write, bash, ls
 ---
 
 You are a test engineer. Identify the right test level, add or adjust tests when asked, detect flaky assumptions, and report exact validation commands and results.

package/agents/writer.md

@@ -5,7 +5,7 @@ model: false
 systemPromptMode: replace
 inheritProjectContext: true
 inheritSkills: false
-tools: read, grep, find, ls, edit, write
+tools: read, edit, write, ls
 ---
 
 You are a documentation specialist. Produce clear, concise, maintainable docs and summaries. Preserve technical accuracy and avoid marketing fluff.

package/dist/index.mjs

@@ -11602,15 +11602,51 @@ var init_frontmatter = __esm({
   }
 });
 
+// src/runtime/role-permission.ts
+function permissionForRole(role) {
+  if (READ_ONLY_ROLES.has(role)) return "read_only";
+  if (WRITE_ROLES.has(role)) return "workspace_write";
+  return "read_only";
+}
+function currentCrewRole(env = process.env) {
+  return env.PI_CREW_ROLE?.trim() || env.PI_TEAMS_ROLE?.trim() || void 0;
+}
+function checkSubagentSpawnPermission(role) {
+  if (!role) return { allowed: true, mode: "workspace_write" };
+  const mode = permissionForRole(role);
+  if (mode === "read_only")
+    return {
+      allowed: false,
+      mode,
+      reason: `Role '${role}' is read-only and cannot spawn additional subagents.`
+    };
+  return { allowed: true, mode };
+}
+var READ_ONLY_ROLES, WRITE_ROLES;
+var init_role_permission = __esm({
+  "src/runtime/role-permission.ts"() {
+    "use strict";
+    READ_ONLY_ROLES = /* @__PURE__ */ new Set(["explorer", "reviewer", "security-reviewer", "analyst", "critic", "planner"]);
+    WRITE_ROLES = /* @__PURE__ */ new Set(["executor", "test-engineer", "writer", "verifier", "agent", "cold-verifier", "chain-executor", "worker"]);
+  }
+});
+
 // src/config/role-tools.ts
 function getToolConfig(role) {
   const key = role.includes("_") ? role.replaceAll("_", "-") : role;
   return ROLE_TOOL_CONFIGS[key] ?? ROLE_TOOL_CONFIGS[role] ?? {};
 }
+function isScratchpadEnabledForRole(role, agent) {
+  const normalized = role.includes("_") ? role.replaceAll("_", "-") : role;
+  if (permissionForRole(normalized) === "read_only") return false;
+  if (agent?.scratchpad === false) return false;
+  return agent?.scratchpad === true || getToolConfig(normalized).scratchpad === true;
+}
 var ROLE_TOOL_CONFIGS;
 var init_role_tools = __esm({
   "src/config/role-tools.ts"() {
     "use strict";
+    init_role_permission();
     ROLE_TOOL_CONFIGS = {
       // Explorer - Read-only exploration; bash is included for git log/show
       // (decisions stream needs commit-history mining) but edit/write stay
@@ -11641,9 +11677,11 @@ var init_role_tools = __esm({
         tools: ["read", "grep", "find", "ls", "glob"],
         excludeTools: ["edit", "write", "bash", "web"]
       },
-      // Executor - Full access (default)
+      // Executor - Full access (default). Phase 1 scratchpad-enabled (stateful
+      // evaluator compounds intermediate results across execute calls).
       executor: {
         // No restrictions - full tool access
+        scratchpad: true
       },
       // Reviewer - Read and review, no write
       reviewer: {
@@ -11670,12 +11708,16 @@ var init_role_tools = __esm({
       // integrity is preserved during verification. Mirrors cold-verifier behavior.
       verifier: {
         tools: ["read", "grep", "find", "ls", "bash"],
-        excludeTools: ["edit", "write", "web"]
+        excludeTools: ["edit", "write", "web"],
+        // Phase 1 scratchpad: multi-cell test/verify flows reuse parsed state.
+        scratchpad: true
       },
       // Test Engineer - Can write tests (F1: hyphenated key)
       "test-engineer": {
         tools: ["read", "edit", "write", "bash", "ls"],
-        excludeTools: ["web"]
+        excludeTools: ["web"],
+        // Phase 1 scratchpad: build/run test suites with state across cells.
+        scratchpad: true
       }
     };
   }
@@ -11891,6 +11933,14 @@ function parseAgentFile(filePath, source) {
       fallbackModels: parseCsv(frontmatter.fallbackModels),
       thinking: frontmatter.thinking === "false" ? void 0 : frontmatter.thinking || void 0,
       tools: parseToolsField(frontmatter.tools),
+      // Phase 1 scratchpad opt-in (Q3: pi ignores unknown frontmatter keys; pi-crew
+      // is the sole consumer in the worker path — task arrives via -p, agent file
+      // is not re-read by pi). 3-STATE parse (NOT `=== "true"` like
+      // inheritProjectContext): omitted/malformed → undefined so the F6 kill-switch
+      // (`agent.scratchpad === false`) only fires on an EXPLICIT `scratchpad: false`,
+      // not on every agent without the key (which would wrongly kill role
+      // default-on). Only the literal "true"/"false" are honored.
+      scratchpad: frontmatter.scratchpad === "true" ? true : frontmatter.scratchpad === "false" ? false : void 0,
       // SEC-1: Strip extensions/excludeExtensions for untrusted project-sourced
       // agents (RCE prevention). Both `project` (.crew/agents/) and
       // `project-pi` (.pi/agents/) are repo-adjacent / untrusted sources —
@@ -14860,6 +14910,55 @@ var init_env_filter = __esm({
   }
 });
 
+// src/runtime/scratchpad/snapshot-lookup.ts
+import { lstatSync as lstatSync5, readdirSync as readdirSync8 } from "node:fs";
+import { join as join18 } from "node:path";
+function findLatestScratchpadSnapshot(artifactsRoot, agentId) {
+  const scratchpadDir = join18(artifactsRoot, "scratchpad");
+  let dirStat;
+  try {
+    dirStat = lstatSync5(scratchpadDir);
+  } catch {
+    return null;
+  }
+  if (dirStat.isSymbolicLink() || !dirStat.isDirectory()) return null;
+  let entries;
+  try {
+    entries = readdirSync8(scratchpadDir, { withFileTypes: true });
+  } catch {
+    return null;
+  }
+  const prefix = `${agentId}.attempt-`;
+  let best = null;
+  for (const dirent of entries) {
+    if (dirent.isSymbolicLink() || !dirent.isFile()) continue;
+    const name = dirent.name;
+    if (!name.startsWith(prefix) || !name.endsWith(SNAPSHOT_SUFFIX)) continue;
+    const attemptPart = name.slice(prefix.length, name.length - SNAPSHOT_SUFFIX.length);
+    if (!/^\d+$/.test(attemptPart)) continue;
+    const attempt = Number.parseInt(attemptPart, 10);
+    let stat2;
+    try {
+      stat2 = lstatSync5(join18(scratchpadDir, name));
+    } catch {
+      continue;
+    }
+    if (!stat2.isFile()) continue;
+    const hit = { path: join18(scratchpadDir, name), attempt, mtimeMs: stat2.mtimeMs };
+    if (best === null || hit.mtimeMs > best.mtimeMs || hit.mtimeMs === best.mtimeMs && hit.attempt < best.attempt) {
+      best = hit;
+    }
+  }
+  return best;
+}
+var SNAPSHOT_SUFFIX;
+var init_snapshot_lookup = __esm({
+  "src/runtime/scratchpad/snapshot-lookup.ts"() {
+    "use strict";
+    SNAPSHOT_SUFFIX = ".snapshot.json";
+  }
+});
+
 // src/runtime/child-pi/child-pi-spawn.ts
 import * as fs18 from "node:fs";
 import * as path18 from "node:path";
@@ -14956,6 +15055,21 @@ function prepareSpawnContext(input, effectiveTask) {
     if (input.runId) built.env.PI_CREW_BROKER_RUN_ID = input.runId;
     if (input.agentId) built.env.PI_CREW_BROKER_TASK_ID = input.agentId;
   }
+  if (input.agentId && isScratchpadEnabledForRole(input.role ?? input.agent.name, input.agent)) {
+    built.env.PI_CREW_SCRATCHPAD = "1";
+    built.env.PI_CREW_TASK_ID = input.agentId;
+    built.env.PI_CREW_ATTEMPT = String(input.attempt ?? 0);
+    if (input.artifactsRoot) {
+      built.env.PI_CREW_ARTIFACTS_ROOT = input.artifactsRoot;
+    }
+    const scratchTempDir = built.tempDir ?? createSafeTempDir(getPiTempBase(), "pi-crew-scratchpad-");
+    built.env.PI_CREW_SCRATCHPAD_SNAPSHOT = resolveRealContainedPath(scratchTempDir, `${input.agentId}.snapshot.json`);
+    const restoreHit = input.artifactsRoot ? findLatestScratchpadSnapshot(input.artifactsRoot, input.agentId) : null;
+    if (restoreHit) {
+      built.env.PI_CREW_SCRATCHPAD_RESTORE = restoreHit.path;
+      built.env.PI_CREW_SCRATCHPAD_RESTORE_MTIME = String(restoreHit.mtimeMs);
+    }
+  }
   if (input.signal?.aborted) {
     return {
       kind: "aborted",
@@ -14983,11 +15097,14 @@ var BASE_ALLOWLIST;
 var init_child_pi_spawn = __esm({
   "src/runtime/child-pi/child-pi-spawn.ts"() {
     "use strict";
+    init_role_tools();
     init_env_allowlist();
     init_env_filter();
     init_internal_error();
+    init_safe_paths();
     init_pi_args();
     init_pi_spawn();
+    init_snapshot_lookup();
     BASE_ALLOWLIST = [
       "PATH",
       "HOME",
@@ -23005,12 +23122,12 @@ var init_crew_hooks = __esm({
 
 // src/runtime/skill-effectiveness.ts
 import { existsSync as existsSync23, mkdirSync as mkdirSync14, readFileSync as readFileSync23, writeFileSync as writeFileSync4 } from "node:fs";
-import { dirname as dirname16, join as join29 } from "node:path";
+import { dirname as dirname16, join as join30 } from "node:path";
 function getSkillMetricsPath(cwd, runId) {
-  return join29(projectCrewRoot(cwd), `state/runs/${runId}/skill-metrics.jsonl`);
+  return join30(projectCrewRoot(cwd), `state/runs/${runId}/skill-metrics.jsonl`);
 }
 function getSkillActivationsPath(cwd, runId) {
-  return join29(projectCrewRoot(cwd), `state/runs/${runId}/skill-activations.jsonl`);
+  return join30(projectCrewRoot(cwd), `state/runs/${runId}/skill-activations.jsonl`);
 }
 function ensureSkillMetricsDir(cwd, runId) {
   const dir = dirname16(getSkillMetricsPath(cwd, runId));
@@ -33743,35 +33860,6 @@ var init_live_session_runtime = __esm({
   }
 });
 
-// src/runtime/role-permission.ts
-function permissionForRole(role) {
-  if (READ_ONLY_ROLES.has(role)) return "read_only";
-  if (WRITE_ROLES.has(role)) return "workspace_write";
-  return "read_only";
-}
-function currentCrewRole(env = process.env) {
-  return env.PI_CREW_ROLE?.trim() || env.PI_TEAMS_ROLE?.trim() || void 0;
-}
-function checkSubagentSpawnPermission(role) {
-  if (!role) return { allowed: true, mode: "workspace_write" };
-  const mode = permissionForRole(role);
-  if (mode === "read_only")
-    return {
-      allowed: false,
-      mode,
-      reason: `Role '${role}' is read-only and cannot spawn additional subagents.`
-    };
-  return { allowed: true, mode };
-}
-var READ_ONLY_ROLES, WRITE_ROLES;
-var init_role_permission = __esm({
-  "src/runtime/role-permission.ts"() {
-    "use strict";
-    READ_ONLY_ROLES = /* @__PURE__ */ new Set(["explorer", "reviewer", "security-reviewer", "analyst", "critic", "planner"]);
-    WRITE_ROLES = /* @__PURE__ */ new Set(["executor", "test-engineer", "writer", "verifier", "agent", "cold-verifier", "chain-executor", "worker"]);
-  }
-});
-
 // src/state/coordination/task-claims.ts
 import { randomUUID as randomUUID4, timingSafeEqual as timingSafeEqual2 } from "node:crypto";
 function createTaskClaim(owner, leaseMs = 5 * 6e4, now = /* @__PURE__ */ new Date()) {
@@ -39426,6 +39514,7 @@ function serializeAgent(agent) {
     line("fallbackModels", agent.fallbackModels),
     line("thinking", agent.thinking),
     line("tools", agent.tools),
+    line("scratchpad", agent.scratchpad),
     agent.extensions !== void 0 ? line("extensions", agent.extensions) ?? "extensions:" : void 0,
     line("skills", agent.skills),
     line("systemPromptMode", agent.systemPromptMode),
@@ -40772,9 +40861,9 @@ var init_handle_settings = __esm({
 
 // src/extension/team-tool/workflow-manage.ts
 import { existsSync as existsSync41, readFileSync as readFileSync40, rmSync as rmSync14, writeFileSync as writeFileSync7 } from "node:fs";
-import { dirname as dirname29, join as join45 } from "node:path";
+import { dirname as dirname29, join as join46 } from "node:path";
 function allowedWorkflowDirs(cwd) {
-  return [join45(projectCrewRoot(cwd), "workflows"), join45(userPiRoot(), "workflows"), join45(packageRoot(), "workflows")];
+  return [join46(projectCrewRoot(cwd), "workflows"), join46(userPiRoot(), "workflows"), join46(packageRoot(), "workflows")];
 }
 function validateScriptContent(content) {
   for (const pattern of FORBIDDEN_PATTERNS) {
@@ -40786,7 +40875,7 @@ function validateScriptContent(content) {
 }
 function resolveWorkflowWritePath(cwd, name, scope = "project") {
   assertSafePathId("workflowName", name);
-  const base = scope === "user" ? join45(userPiRoot(), "workflows") : join45(projectCrewRoot(cwd), "workflows");
+  const base = scope === "user" ? join46(userPiRoot(), "workflows") : join46(projectCrewRoot(cwd), "workflows");
   return resolveRealContainedPath(base, `${name}.dwf.ts`);
 }
 function handleWorkflowCreate(params, ctx) {
@@ -41640,7 +41729,7 @@ var init_async_runner = __esm({
 });
 
 // src/runtime/goal-workflow/goal-state-store.ts
-import { closeSync as closeSync10, existsSync as existsSync44, mkdirSync as mkdirSync26, openSync as openSync10, readdirSync as readdirSync20, readFileSync as readFileSync42, statSync as statSync33, unlinkSync as unlinkSync7 } from "node:fs";
+import { closeSync as closeSync10, existsSync as existsSync44, mkdirSync as mkdirSync26, openSync as openSync10, readdirSync as readdirSync21, readFileSync as readFileSync42, statSync as statSync33, unlinkSync as unlinkSync7 } from "node:fs";
 import { dirname as dirname32 } from "node:path";
 function resolveGoalsRoot(cwd) {
   const crewRoot = projectCrewRoot(cwd) ?? userCrewRoot();
@@ -41800,7 +41889,7 @@ var init_goal_state_store = __esm({
         try {
           const root = resolveGoalsRoot(this.cwd);
           if (!existsSync44(root)) return [];
-          const entries = readdirSync20(root);
+          const entries = readdirSync21(root);
           const goals = [];
           for (const entry of entries) {
             if (!entry.endsWith(".json")) continue;
@@ -41864,7 +41953,7 @@ var init_verification_integrity = __esm({
 
 // src/runtime/workspace-lock.ts
 import { createHash as createHash7 } from "node:crypto";
-import { closeSync as closeSync11, existsSync as existsSync45, mkdirSync as mkdirSync27, openSync as openSync11, readdirSync as readdirSync21, readFileSync as readFileSync44, statSync as statSync35, unlinkSync as unlinkSync8, writeFileSync as writeFileSync8 } from "node:fs";
+import { closeSync as closeSync11, existsSync as existsSync45, mkdirSync as mkdirSync27, openSync as openSync11, readdirSync as readdirSync22, readFileSync as readFileSync44, statSync as statSync35, unlinkSync as unlinkSync8, writeFileSync as writeFileSync8 } from "node:fs";
 import * as path50 from "node:path";
 function workspaceLockPath(cwd) {
   const absCwd = path50.resolve(cwd);
@@ -45069,7 +45158,8 @@ __export(crash_recovery_exports, {
   detectInterruptedRuns: () => detectInterruptedRuns,
   purgeStaleActiveRunIndex: () => purgeStaleActiveRunIndex,
   readManifestWithTransientRetry: () => readManifestWithTransientRetry,
-  reconcileAllStaleRuns: () => reconcileAllStaleRuns
+  reconcileAllStaleRuns: () => reconcileAllStaleRuns,
+  shouldRecoverTask: () => shouldRecoverTask
 });
 import * as fs63 from "node:fs";
 import * as path52 from "node:path";
@@ -53188,6 +53278,20 @@ function detectRetryableModelFailureFromOutput(parsed) {
   }
   return void 0;
 }
+function evidenceStatusFor(childResult) {
+  return childResult.exitStatus?.cancelled ? "cancelled" : childResult.error || childResult.exitCode && childResult.exitCode !== 0 ? "failed" : "completed";
+}
+function attemptErrorFor(childResult, parsedOutput, taskId) {
+  let err2 = childResult.error || (childResult.exitCode && childResult.exitCode !== 0 ? childResult.stderr || `Child Pi exited with ${childResult.exitCode}` : void 0);
+  if (childResult.exitStatus?.timedOut) {
+    err2 = errors.childTimeout({ taskId, stderr: childResult.stderr }).message;
+  }
+  if (!err2 && parsedOutput) {
+    const rateLimitErr = detectRetryableModelFailureFromOutput(parsedOutput);
+    if (rateLimitErr) err2 = rateLimitErr;
+  }
+  return err2;
+}
 async function runChildProcessTask(ctx) {
   const input = ctx.input;
   const manifest = ctx.manifest;
@@ -53371,6 +53475,7 @@ async function runChildProcessTask(ctx) {
         runId: manifest.runId,
         agentId: task.id,
         artifactsRoot: manifest.artifactsRoot,
+        attempt: i,
         steeringFile: resolveRealContainedPath(`${manifest.artifactsRoot}/steering`, `${task.id}.jsonl`),
         onSpawn: (pid) => {
           try {
@@ -53459,7 +53564,7 @@ async function runChildProcessTask(ctx) {
         input.signal.removeEventListener("abort", externalAbortListener);
       }
     }
-    const evidenceStatus = childResult.exitStatus?.cancelled ? "cancelled" : childResult.error || childResult.exitCode && childResult.exitCode !== 0 ? "failed" : "completed";
+    const evidenceStatus = evidenceStatusFor(childResult);
     terminalEvidence = [
       ...terminalEvidence,
       {
@@ -53505,17 +53610,7 @@ async function runChildProcessTask(ctx) {
     parsedOutput = parsePiJsonOutput(transcriptText2);
     rawFinalText = childResult.rawFinalText;
     intermediateFindings = childResult.intermediateFindings;
-    error = childResult.error || (childResult.exitCode && childResult.exitCode !== 0 ? childResult.stderr || `Child Pi exited with ${childResult.exitCode}` : void 0);
-    if (childResult.exitStatus?.timedOut) {
-      error = errors.childTimeout({
-        taskId: task.id,
-        stderr: childResult.stderr
-      }).message;
-    }
-    if (!error && parsedOutput) {
-      const rateLimitErr = detectRetryableModelFailureFromOutput(parsedOutput);
-      if (rateLimitErr) error = rateLimitErr;
-    }
+    error = attemptErrorFor(childResult, parsedOutput, task.id);
     persistHeartbeat(true);
     persistChildProgress({ type: "attempt_finished" }, true);
     const attempt = {
@@ -61073,7 +61168,7 @@ __export(dynamic_workflow_runner_exports, {
   runDynamicWorkflow: () => runDynamicWorkflow
 });
 import { readFileSync as readFileSync67 } from "node:fs";
-import { join as join71 } from "node:path";
+import { join as join72 } from "node:path";
 import { transformSync } from "esbuild";
 function assertStructuredCloneable(value, name) {
   try {
@@ -61085,7 +61180,7 @@ function assertStructuredCloneable(value, name) {
 }
 function resolveScriptPath(workflow, cwd) {
   const crewRoot = projectCrewRoot(cwd);
-  const allowedBases = [join71(projectCrewRoot(cwd), "workflows"), join71(userPiRoot(), "workflows"), join71(packageRoot(), "workflows")];
+  const allowedBases = [join72(projectCrewRoot(cwd), "workflows"), join72(userPiRoot(), "workflows"), join72(packageRoot(), "workflows")];
   for (const base of allowedBases) {
     try {
       const real = resolveRealContainedPath(base, workflow.filePath);
@@ -72238,20 +72333,20 @@ init_internal_error();
 
 // src/extension/crew-vibes/config.ts
 import { existsSync as existsSync75, mkdirSync as mkdirSync42, readFileSync as readFileSync74, writeFileSync as writeFileSync9 } from "node:fs";
-import { dirname as dirname39, join as join77 } from "node:path";
+import { dirname as dirname39, join as join78 } from "node:path";
 
 // src/extension/crew-vibes/font-detect.ts
 import { existsSync as existsSync74, readFileSync as readFileSync73 } from "node:fs";
 import { homedir as homedir11, platform } from "node:os";
-import { join as join76 } from "node:path";
+import { join as join77 } from "node:path";
 function fontPath() {
   const os18 = platform();
   const home = homedir11();
-  if (os18 === "darwin") return join76(home, "Library", "Fonts", "crew-vibes.ttf");
-  if (os18 === "linux") return join76(home, ".local", "share", "fonts", "crew-vibes.ttf");
+  if (os18 === "darwin") return join77(home, "Library", "Fonts", "crew-vibes.ttf");
+  if (os18 === "linux") return join77(home, ".local", "share", "fonts", "crew-vibes.ttf");
   if (os18 === "win32") {
-    const local = process.env.LOCALAPPDATA ?? join76(home, "AppData", "Local");
-    return join76(local, "Microsoft", "Windows", "Fonts", "crew-vibes.ttf");
+    const local = process.env.LOCALAPPDATA ?? join77(home, "AppData", "Local");
+    return join77(local, "Microsoft", "Windows", "Fonts", "crew-vibes.ttf");
   }
   return "";
 }
@@ -72300,7 +72395,7 @@ function resolveHome() {
   return (process.env.PI_TEAMS_HOME ?? process.env.PI_CREW_HOME)?.trim() || process.env.HOME || process.env.USERPROFILE || "";
 }
 function configPath2() {
-  return join77(resolveHome(), ".pi", "agent", "pi-crew-vibes.json");
+  return join78(resolveHome(), ".pi", "agent", "pi-crew-vibes.json");
 }
 var DEFAULT_CONFIG2 = {
   enabled: true,
@@ -72800,14 +72895,14 @@ function createCrewVibesFooter(deps) {
 // src/extension/crew-vibes/provider-usage.ts
 import { readFileSync as readFileSync75 } from "node:fs";
 import { homedir as homedir12 } from "node:os";
-import { join as join78 } from "node:path";
+import { join as join79 } from "node:path";
 function withTimeout(ms, fn) {
   const controller = new AbortController();
   const timeoutId = setTimeout(() => controller.abort(), ms);
   return fn(controller.signal).finally(() => clearTimeout(timeoutId));
 }
 function piAuthPath() {
-  return join78(homedir12(), ".pi", "agent", "auth.json");
+  return join79(homedir12(), ".pi", "agent", "auth.json");
 }
 function loadAnthropicToken() {
   const envToken = process.env.ANTHROPIC_OAUTH_TOKEN?.trim();
@@ -72852,8 +72947,8 @@ function tokenFromHostEntry(entry) {
   return void 0;
 }
 function loadLegacyCopilotToken() {
-  const configHome = process.env.XDG_CONFIG_HOME?.trim() || join78(homedir12(), ".config");
-  const candidates = [join78(configHome, "github-copilot", "hosts.json"), join78(homedir12(), ".github-copilot", "hosts.json")];
+  const configHome = process.env.XDG_CONFIG_HOME?.trim() || join79(homedir12(), ".config");
+  const candidates = [join79(configHome, "github-copilot", "hosts.json"), join79(homedir12(), ".github-copilot", "hosts.json")];
   for (const hostsPath of candidates) {
     try {
       const data = JSON.parse(readFileSync75(hostsPath, "utf8"));

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pi-crew",
-  "version": "0.9.62",
+  "version": "0.9.64",
   "description": "Pi extension for coordinated AI teams, workflows, worktrees, and async task orchestration",
   "author": "baphuongna",
   "license": "MIT",
@@ -61,7 +61,9 @@
     "types/",
     "CHANGELOG.md",
     "LICENSE",
-    "NOTICE.md"
+    "NOTICE.md",
+    "scripts/resource-sampler.mjs",
+    "scripts/analyze-run.mjs"
   ],
   "scripts": {
     "check": "npm run ci",
@@ -79,6 +81,7 @@
     "test:watch": "tsx --watch --test --test-concurrency=4 --test-timeout=30000 --test-force-exit 'test/unit/**/*.test.ts'",
     "test:integration": "node scripts/test-runner.mjs --test-concurrency=1 --test-timeout=300000 test/integration/*.test.ts",
     "test:smoke": "node scripts/test-runner.mjs --test-concurrency=1 --test-timeout=180000 test/smoke/*.smoke.ts",
+    "test:spike": "node scripts/test-runner.mjs --test-concurrency=2 --test-timeout=120000 --test-force-exit test/runtime/scratchpad/*.test.ts",
     "build:bundle": "node scripts/build-bundle.mjs",
     "test:bundle": "node --experimental-strip-types --test --test-force-exit test/unit/bundle-load.test.ts",
     "watch:bundle": "node scripts/watch-bundle.mjs",