pi-crew 0.9.51 → 0.9.53

package/docs/TEST_MATRIX.md

@@ -1,51 +0,0 @@
-# Test Matrix
-
-Maps pi-crew behavior to proof. Every row must have real validation evidence.
-
-## Status Values
-
-| Status | Meaning |
-|--------|---------|
-| planned | Accepted behavior, not implemented |
-| in_progress | Actively being built |
-| implemented | Implemented and proof exists |
-| changed | Contract changed after implementation |
-| retired | No longer part of product |
-
-## Matrix
-
-| Story | Contract | Unit | Integration | CI | Status | Evidence |
-|-------|----------|------|-------------|-----|--------|----------|
-| Core team run | `docs/product/team-run.md` | yes | yes | yes 3/3 | implemented | 2703 tests pass (133 suites) |
-| Child process runner | `docs/product/child-process.md` | yes | yes | yes 3/3 | implemented | child-pi-pool.test.ts, child-pi-timeout.test.ts, mock-child-run.test.ts |
-| Async runner | `docs/product/async-runner.md` | yes | yes | yes 3/3 | implemented | async-runner.test.ts, async-restart-recovery.test.ts |
-| Live session | `docs/product/live-session.md` | yes | no | yes 3/3 | implemented | live-session-context.test.ts, live-session-runtime.test.ts |
-| State durability | `docs/product/state.md` | yes | yes | yes 3/3 | implemented | state-store.test.ts, state-contracts.test.ts, phase3-runtime.test.ts |
-| Worktree isolation | `docs/product/worktree.md` | yes | yes | yes 3/3 | implemented | worktree-manager.test.ts, worktree-run.test.ts |
-| Team tool API | `docs/product/team-tool.md` | yes | yes | yes 3/3 | implemented | team-tool-dispatch.test.ts, extension-api-surface.test.ts, operator-experience.test.ts |
-| Group join | `docs/product/group-join.md` | yes | yes | yes 3/3 | implemented | phase6-runtime-hardening.test.ts |
-| Model fallback | `docs/product/model-fallback.md` | yes | no | yes 3/3 | implemented | model-fallback.test.ts |
-| Conflict detection | `docs/product/conflict-detect.md` | yes | no | yes 3/3 | implemented | conflict-detect.test.ts, delta-conflict.test.ts |
-| Crash recovery | `docs/product/crash-recovery.md` | yes | yes | yes 3/3 | implemented | recovery-recipes.test.ts, async-restart-recovery.test.ts |
-| Effectiveness guard | `docs/product/effectiveness.md` | yes | no | yes 3/3 | implemented | effectiveness-guard.test.ts |
-| Windows EBUSY | `docs/product/platform.md` | yes | yes | yes 3/3 | implemented | phase6-runtime-hardening.test.ts |
-| Depth guard | `docs/product/runtime-safety.md` | yes | no | yes 3/3 | implemented | subagent-depth.test.ts, completion-guard.test.ts |
-
-## Evidence Rules
-
-- **Unit proof**: Pure logic, state transitions, config parsing
-- **Integration proof**: Multi-module interaction (team runner → state → child process)
-- **CI proof**: Cross-platform (ubuntu, windows, macos) green on GitHub Actions
-- A story can be implemented without every proof column if the story explains why
-- Agents must run `npm test` and `npm run typecheck` before claiming done
-
-## Validation Commands
-
-```bash
-npm test                    # Run all unit tests (1655 tests across 268 unit files + 14 integration files)
-npm run typecheck           # TypeScript check + strip-types import
-npm run check               # Biome lint + format
-npm run test:unit           # Unit tests only (fast, parallel)
-npm run test:integration    # Integration tests only (sequential)
-gh run list --limit 1       # Check latest CI status
-```

package/docs/archive/README.md

@@ -1,91 +0,0 @@
-# Archived Documents
-
-This directory contains historical one-off reports and reviews that were
-moved out of the main `docs/` directory as part of the **L3 (docs sprawl
-cleanup)** in the v0.9.17 cleanup plan.
-
-**Criterion (per plan §11 L3 circle-back):** "exclude any file referenced by
-AGENTS.md or README." Each file here has been verified to have NO source-of-
-truth reference in the project — these are point-in-time audits, reviews,
-test-round logs, and followup reports that were used once and superseded.
-
-## What is archived (22 files)
-
-### v0.5.x audit-fix plans (9 files)
-Spans the v0.5.5 → v0.5.17 release cycle of bug audits. Each was a point-in-time
-list of fixes for that release; all fixes were shipped and the audit is now
-historical.
-
-- `pi-crew-v0.5.5-audit-fix-plan.md`
-- `pi-crew-v0.5.9-audit-fix-plan.md`
-- `pi-crew-v0.5.10-audit-fix-plan.md`
-- `pi-crew-v0.5.11-audit-fix-plan.md`
-- `pi-crew-v0.5.12-audit-fix-plan.md`
-- `pi-crew-v0.5.13-audit-fix-plan.md`
-- `pi-crew-v0.5.14-audit-fix-plan.md`
-- `pi-crew-v0.5.16-audit-fix-plan.md`
-- `pi-crew-v0.5.17-audit-fix-plan.md`
-
-### Round-series test logs (6 files)
-The iterative test result logs from `pi-crew-test-roundN` rounds. v0.9.17 has
-a new test matrix (`docs/TEST_MATRIX.md`) which supersedes these.
-
-- `pi-crew-test-final.md`
-- `pi-crew-test-results.md`
-- `pi-crew-test-round2.md`
-- `pi-crew-test-round4.md`
-- `pi-crew-test-round5.md`
-- `pi-crew-test-round6.md`
-
-### Review-round followups (2 files)
-Sequential review rounds from May 2026.
-
-- `followup-review-round3-2026-05-12.md`
-- `followup-review-round4-2026-05-13.md`
-
-### Single-investigation reports (5 files)
-One-off reports retained for historical reference only.
-
-- `pi-crew-bugs.md` — early-stage bug inventory (May 2026)
-- `pi-mono-opportunities.md` — cross-project opportunity scan
-- `pi-mono-review.md` — cross-project review
-- `pi-subagent4-comparison.md` — subagent-4 comparison
-- `pi-subagents3-deep-analysis.md` — subagents-3 deep analysis
-
-## How to recover a file
-
-Files in this directory are still tracked by git (history preserved). To
-restore:
-
-```bash
-git mv docs/archive/FILE.md docs/
-# edit docs/README.md or other references if needed
-git commit -m "docs: restore FILE.md from archive"
-```
-
-## Why these were archived
-
-`docs/` had grown to ~77 markdown files. Many were one-off reports created
-during specific development cycles that:
-
-- had no source-of-truth reference (not linked from README, AGENTS.md, or
-  any code file's header comment),
-- were superseded by current docs (e.g. `TEST_MATRIX.md`),
-- were point-in-time audits that produced code changes already shipped.
-
-This directory keeps them available for git-history archaeology without
-cluttering the active docs surface.
-
-## What is NOT archived (still in `docs/`)
-
-- `architecture.md`, `actions-reference.md`, `commands-reference.md`,
-  `dynamic-workflows.md` — referenced from `README.md`
-- `decisions/` (ADR-style 0001-0008) — referenced from `AGENTS.md`
-- `FEATURE_INTAKE.md`, `HARNESS.md`, `HARNESS_BACKLOG.md`, `TEST_MATRIX.md`,
-  `troubleshooting.md`, `usage.md`, `goals.md`, `live-mailbox-runtime.md`,
-  `runtime-flow.md`, `resource-formats.md`, `publishing.md`,
-  `migration-v0.4-v0.5.md` — active project docs
-- `bugs/`, `fixes/`, `distillation/`, `product/`, `stories/`, `skills/`,
-  `templates/`, `superpowers/`, `perf/`, `patterns/` — themed subdirs
-- `next-upgrade-roadmap.md` — referenced from `README.md`
-- `REVIEW-FINDINGS-2026-06` (file) — current cleanup plan reference

package/docs/archive/followup-review-round3-2026-05-12.md

@@ -1,342 +0,0 @@
-# Follow-up Review — pi-crew (2026-05-12, round 3)
-
-Author: Droid (Factory) | Related: `docs/code-review-2026-05-11.md`, `docs/followup-plan-2026-05-12.md`, `docs/followup-review-2026-05-12.md`. HEAD: `5bee878`.
-
-This is the review round after commit `5bee878` resolved C1–C6. Goal: scrutinize the modules not yet examined closely (event-log, atomic-write, child-pi, redaction, sleep, hooks, cleanup) to find remaining risks.
-
-## Result summary
-
-- `npm run typecheck` → Passed
-- `npm run check:lazy-imports` → Passed
-- `npm run test:unit` → **1418 tests / 1415 pass / 0 fail / 3 skip** (212s)
-
-The codebase is in a stable state. The findings below are **low risk** or **defense-in-depth**, not urgent bugs.
-
----
-
-## Part A — New findings
-
-### D1 — `event-log.appendEvent` has no lock, JSONL may interleave on Windows
-
-**Severity:** Medium | **Effort:** ~30 minutes | **File:** `src/state/event-log.ts:148`
-
-**Current state:**
-```ts
-fs.appendFileSync(eventsPath, `${JSON.stringify(redactSecrets(fullEvent))}\n`, "utf-8");
-```
-
-**Problem:**
-- `fs.appendFileSync` on POSIX is only atomic for writes smaller than `PIPE_BUF` (~4 KiB). A full event JSON (with data, metadata, transcripts) can exceed that → interleaved lines between 2 processes (parent + background-runner).
-- On Windows, append is NOT atomic at any size; 2 processes appending to the same eventsPath can produce interleaved JSON lines → `JSON.parse(line)` in `readEvents`/`scanSequence` throws and skips the line.
-- Consequence: lost events, sequence numbers jumping, "appended: false" is returned in a different path (size-limit) but the normal path gives no hint.
-
-**Trigger:** running `background-runner` in parallel with the parent writing events to the same `eventsPath` (e.g. cancel + retry in succession).
-
-**Proposed fix:**
-1. Wrap `appendEvent` in `withRunLockSync(manifest, () => { ... })` — ensures exclusive access.
-2. Or use `fs.openSync(..., O_APPEND | O_WRONLY)` + retry with an advisory lock (`flock` POSIX, `LockFileEx` Windows — via the npm package `proper-lockfile`).
-3. Lightest option: switch to `appendEventAsync` via a queue/serialize.
-
-**Tests to add:** stress test in `test/integration/`: 2 processes each appending 100 events concurrently → assert total parseable line count = 200.
-
----
-
-### D2 — `event-log.sequenceCache` Map leaks by number of runs
-
-**Severity:** Low | **Effort:** ~10 minutes | **File:** `src/state/event-log.ts:60`
-
-**Current state:**
-```ts
-const sequenceCache = new Map<string, { size: number; mtimeMs: number; seq: number }>();
-```
-
-**Problem:**
-- Module-level map, never evicts. Each `eventsPath` (1 per run) takes 1 entry. A long-running parent process (e.g. live-session-runtime) running for many days → the cache could reach thousands of entries.
-- Memory isn't large (~100 bytes/entry) but is unbounded.
-
-**Proposed fix:** use a simple LRU (Map with a max size, evict oldest when exceeding the threshold), or clear after the run ends:
-```ts
-export function evictSequenceCache(eventsPath: string): void {
-	sequenceCache.delete(eventsPath);
-}
-// Call from updateRunStatus(..., "completed"/"failed"/"cancelled").
-```
-
----
-
-### D3 — `atomicWriteFileAsync` has a "matches" fallback → sync path doesn't (parity)
-
-**Severity:** Low | **Effort:** ~15 minutes | **File:** `src/state/atomic-write.ts:122-138`
-
-**Current state:**
-```ts
-// async path:
-try { await renameWithRetryAsync(...); }
-catch (renameError) {
-	const existing = await fs.promises.readFile(filePath, "utf-8");
-	const matches = existing === content;
-	if (matches) { /* cleanup temp, return success */ }
-	throw renameError;
-}
-
-// sync path: just throws, no "matches" fallback.
-```
-
-**Problem:**
-- The async path "forgives" the race condition (the file was already written with the correct content by another process). The sync path throws hard.
-- Different semantics → hard to debug when someone uses sync with a race.
-- This case is rare (identical content), but the asymmetry is a code smell.
-
-**Proposed fix:** add the same fallback to sync, or remove the fallback from async (pick one consistent convention):
-```ts
-} catch (renameError) {
-	try {
-		const existing = fs.readFileSync(filePath, "utf-8");
-		if (existing === content) {
-			try { fs.rmSync(tempPath, { force: true }); } catch { /* best-effort */ }
-			return;
-		}
-	} catch { /* fall through */ }
-	throw renameError;
-}
-```
-
----
-
-### D4 — `withRunLock` (async) waits until the deadline for an active lock, `withRunLockSync` throws immediately
-
-**Severity:** Low | **Effort:** ~10 minutes | **File:** `src/state/locks.ts:91-110`
-
-**Current state:**
-- Sync: `if (!isLockStale(...)) throw ...` → fail fast for an active lock.
-- Async: only checks staleness in `readLockStateAsync`, doesn't throw for active → loops waiting until the deadline (`staleMs * 2`, usually 60s).
-
-**Problem:**
-- The test `withRunLockSync throws immediately on active (non-stale) lock` proves sync throws immediately.
-- Async will hang ~60s then throw → slow cancel/retry experience.
-- BUG-004 (round 1) aimed to unify sync ↔ async, but this semantic asymmetry remains.
-
-**Proposed fix:** unify by one of:
-- Sync: add a short wait + retry like async (wait up to 1-2s then throw).
-- Async: throw immediately when the lock isn't stale (like sync) — usually better since the caller can retry with higher context.
-
-```ts
-async function acquireLockWithRetryAsync(filePath: string, staleMs: number): Promise<void> {
-	let attempt = 0;
-	const deadline = Date.now() + staleMs * 2;
-	while (true) {
-		try { writeLockFile(filePath); return; }
-		catch (error) {
-			const code = (error as NodeJS.ErrnoException).code;
-			if (code !== "EEXIST") throw error;
-			if (Date.now() > deadline) throw new Error(`Run '${path.basename(filePath)}' is locked.`);
-			if (!isLockStale(filePath, staleMs)) {
-				throw new Error(`Run '${path.basename(filePath)}' is locked by another operation.`);
-			}
-			try { fs.rmSync(filePath, { force: true }); } catch { /* race */ }
-			await sleep(Math.min(250, 25 * 2 ** attempt));
-			attempt++;
-		}
-	}
-}
-```
-
-**Tests to add:** mirror the sync test — `withRunLock async throws immediately on active (non-stale) lock`.
-
----
-
-### D5 — `sleep.ts` uses `require()` in an ES module
-
-**Severity:** Low (style) | **Effort:** ~5 minutes | **File:** `src/utils/sleep.ts:18`
-
-**Current state:**
-```ts
-const { execFileSync } = require("node:child_process") as typeof import("node:child_process");
-```
-
-**Problem:**
-- The project is ESM (`"type": "module"`). `require` only works via strip-types backward-compat — not clean.
-- AGENTS.md: "Avoid dynamic inline imports, EXCEPT at documented lazy-load boundaries to defer heavy runtime cost (mark with `// LAZY: <reason>`)". This `require` has no marker.
-- `child_process` isn't heavy — top-level import is fine.
-
-**Proposed fix:**
-```ts
-import { execFileSync } from "node:child_process";
-// ...
-execFileSync("sleep", [(ms / 1000).toFixed(3)], { timeout: ms + 1000, stdio: "pipe" });
-```
-
----
-
-### D6 — `iteration-hooks.runIterationHook` doesn't filter env like post-checks
-
-**Severity:** Low | **Effort:** ~5 minutes | **File:** `src/runtime/iteration-hooks.ts:140`
-
-**Current state:**
-```ts
-env: { PATH: process.env.PATH ?? "/usr/bin:/bin", HOME: process.env.HOME ?? "/tmp", USER: process.env.USER, LANG: process.env.LANG, PI_CREW_HOOK: "1" },
-```
-
-**Problem:**
-- Already restricted manually, OK for Linux. But on Windows it lacks `USERPROFILE`, `TEMP`, `TMP`, `ComSpec`, `SystemRoot` → `.cmd/.ps1` scripts may fail.
-- post-checks.ts has the same pattern (line 82) — inconsistent with worktree-manager.runSetupHook which moved to `sanitizeEnvSecrets(..., { allowList: [...] })`.
-
-**Proposed fix:** apply `sanitizeEnvSecrets` with an allowList, unifying all 3 sites (post-checks, iteration-hooks, setup-hook):
-```ts
-import { sanitizeEnvSecrets } from "../utils/env-filter.ts";
-const HOOK_ENV_ALLOW = ["PATH", "HOME", "USER", "USERPROFILE", "TEMP", "TMP", "TMPDIR", "LANG", "LC_ALL", "ComSpec", "SystemRoot", "PI_*"];
-// ...
-env: { ...sanitizeEnvSecrets(process.env, { allowList: HOOK_ENV_ALLOW }), PI_CREW_HOOK: "1" },
-```
-
-**Benefits:**
-- 1 source of truth for the hook env whitelist.
-- Supports Windows .cmd/.ps1 (USERPROFILE/TEMP needed).
-- Avoids code duplication.
-
----
-
-### D7 — `cleanup.ts` git helper doesn't force locale (consistency with worktree-manager)
-
-**Severity:** Info | **Effort:** ~2 minutes | **File:** `src/worktree/cleanup.ts:15`
-
-**Current state:**
-```ts
-function git(cwd: string, args: string[]): string {
-	return execFileSync("git", args, { cwd, encoding: "utf-8", stdio: ["ignore", "pipe", "pipe"] }).trim();
-}
-```
-
-**Problem:**
-- `worktree-manager.ts` already forces `LANG: "C", LC_ALL: "C"` (C5 round 2). `cleanup.ts` doesn't yet.
-- Doesn't cause a current bug because cleanup doesn't parse error strings; but if error parsing is added later, this gets missed again.
-- `branch-freshness.ts` has the same issue.
-
-**Proposed fix:** extract a `git()` helper into a shared `src/utils/git-helper.ts`, used by all 3 files:
-```ts
-// src/utils/git-exec.ts
-import { execFileSync } from "node:child_process";
-export function gitExec(cwd: string, args: string[]): string {
-	return execFileSync("git", args, {
-		cwd, encoding: "utf-8", stdio: ["ignore", "pipe", "pipe"],
-		env: { ...process.env, LANG: "C", LC_ALL: "C" },
-	}).trim();
-}
-```
-
----
-
-### D8 — `redaction.PEM_PRIVATE_KEY_PATTERN` has no length limit → low ReDoS potential
-
-**Severity:** Info | **Effort:** ~5 minutes | **File:** `src/utils/redaction.ts:7`
-
-**Current state:**
-```ts
-const PEM_PRIVATE_KEY_PATTERN = /-----BEGIN [A-Z ]*PRIVATE KEY-----[\s\S]*?-----END [A-Z ]*PRIVATE KEY-----/g;
-```
-
-**Problem:**
-- The lazy `[\s\S]*?` is ReDoS-safe, but if the input has a BEGIN without an END → backtracks to the end of the string. With a long JSONL transcript (10+ MB), the regex scans the whole thing.
-- Not a true ReDoS (linear), but can be slow.
-
-**Proposed fix:** add a hard 64KB limit for a PEM block (real PEM ~3KB):
-```ts
-const PEM_PRIVATE_KEY_PATTERN = /-----BEGIN [A-Z ]*PRIVATE KEY-----[\s\S]{0,65536}?-----END [A-Z ]*PRIVATE KEY-----/g;
-```
-
-Trade-off: PEM > 64KB won't be fully redacted. Rare in practice.
-
----
-
-### D9 — `subagent-manager.persistedSubagentPath` doesn't validate `id` → potential path traversal
-
-**Severity:** Low | **Effort:** ~5 minutes | **File:** `src/runtime/subagent-manager.ts:58`
-
-**Current state:**
-```ts
-function persistedSubagentPath(cwd: string, id: string): string {
-	return path.join(projectCrewRoot(cwd), DEFAULT_PATHS.state.subagentsSubdir, `${id}.json`);
-}
-```
-
-**Problem:**
-- `id` is currently generated internally (`agent_${Date.now().toString(36)}_${counter.toString(36)}`) → safe.
-- But `readPersistedSubagentRecord(cwd, id)` is called with `id` from an external source (e.g. `get_subagent_result` tool param). If tool param validation is missing, `id = "../../../etc/passwd"` could read a file outside the state dir.
-
-**Proposed fix:** validate `id` matches `^[a-z0-9_]+$`:
-```ts
-function isValidSubagentId(id: string): boolean {
-	return /^[a-z0-9_]+$/i.test(id) && id.length <= 128;
-}
-function persistedSubagentPath(cwd: string, id: string): string {
-	if (!isValidSubagentId(id)) throw new Error(`Invalid subagent id: ${id}`);
-	return path.join(projectCrewRoot(cwd), DEFAULT_PATHS.state.subagentsSubdir, `${id}.json`);
-}
-```
-
-Check the `get_subagent_result` tool schema to see if it sanitizes id already; if so, D9 is just defense-in-depth.
-
----
-
-## Implementation priority
-
-| # | Item | Severity | Effort | Recommendation |
-|---|---|---|---|---|
-| 1 | D1 (event-log concurrent append) | Medium | 30 minutes | Current sprint — prevent event loss/corruption |
-| 2 | D6 (consistent hook env allowList) | Low | 5 minutes | Current sprint — sync with the setup-hook fix |
-| 3 | D4 (async lock fail-fast for active) | Low | 10 minutes | Current sprint — cancel/retry UX |
-| 4 | D9 (subagent id validate) | Low | 5 minutes | Current sprint — defense-in-depth |
-| 5 | D2 (sequenceCache eviction) | Low | 10 minutes | Next sprint |
-| 6 | D3 (atomic-write sync/async parity) | Low | 15 minutes | Next sprint |
-| 7 | D5 (sleep.ts ESM require) | Low | 5 minutes | Next sprint |
-| 8 | D7 (git helper consolidate) | Info | 2 minutes | Anytime |
-| 9 | D8 (PEM regex limit) | Info | 5 minutes | Anytime |
-
-**Total effort priority 1 (must-fix):** ~50 minutes.
-**Total effort priority 2 (nice-to-have):** ~37 minutes.
-
----
-
-## Proposed commit batches
-
-- **Batch 1 (correctness/security):** D1 + D9 + D4 → 1 PR "event-log lock + subagent id guard + async lock parity" (~45 minutes).
-- **Batch 2 (hardening):** D6 + D7 + D2 → 1 PR "hook env allowList consolidation + git helper extract + cache eviction" (~20 minutes).
-- **Batch 3 (polish):** D3 + D5 + D8 → 1 PR "atomic-write parity + ESM cleanup + redaction limit" (~25 minutes).
-
----
-
-## Positive notes after round 3
-
-- All C1–C6 (round 2) fixed correctly per spec.
-- 1418 tests pass (vs 1411 the previous round → +7 new tests), 0 fail.
-- `npm run check:lazy-imports` now runs on Windows (after removing `sed`).
-- `sanitizeEnvSecrets` has both a deny-list (default) and an allow-list mode → good flexibility.
-- `resolveShellForScript` correctly handles `.bat/.cmd` to guard against CVE-2024-27980.
-- `parent-guard` polling works well cross-platform (POSIX + Windows).
-- Multi-layer redaction pipeline (key-name + inline-substring + auth-header + bearer + PEM).
-- Atomic-write has O_EXCL + O_NOFOLLOW + post-open `isFile()` verification.
-- Subagent records persisted under the redaction filter.
-- Background runner has `parent-guard` + tempdir cleanup + final-drain timer.
-
----
-
-## Areas with NO serious issues (reviewed)
-
-- `src/schema/team-tool-schema.ts` — TypeBox schema has the "retry" literal, strict additionalProperties.
-- `src/state/artifact-store.ts` — 2-layer path traversal blocking (`resolveInside` + `resolveRealContainedPath`), post-redaction hash.
-- `src/state/atomic-write.ts` — symlink-safe, O_EXCL, fd-based stat verification.
-- `src/worktree/worktree-manager.ts` — branchExists local+remote, prune stale, env filter, locale-safe error parse.
-- `src/runtime/async-runner.ts` — jiti + strip-types fallback, multi-candidate path.
-- `src/runtime/child-pi.ts` — env sanitize, redacted transcript, post-exit stdio guard, hard kill timer.
-- `src/runtime/parent-guard.ts` — kill(pid,0) cross-platform, unref'd interval.
-
----
-
-## Verification
-
-```
-npm run typecheck            → Passed
-npm run check:lazy-imports   → All dynamic imports have `// LAZY:` marker.
-npm run test:unit            → 1418 tests, 1415 pass, 0 fail, 3 skip (212s)
-HEAD                         → 5bee878 (fix: address followup-review C1-C6)
-```

package/docs/archive/followup-review-round4-2026-05-13.md

@@ -1,107 +0,0 @@
-# Follow-up Review Round 4 — 2026-05-13
-
-Review of commits `c7bd455` through `5f47e92` (7 commits on top of `faa81e4`).
-
-## Summary
-
-These commits harden live-session runtime cleanup, async runner lifecycle, crew-agent persistence, and Windows test flakiness. Overall quality is high and the fixes address real production issues. However, one commit introduces a **regression** that breaks the resume/follow-up capability for completed live-session agents. Two other issues could cause crashes or races in edge cases.
-
-**All identified bugs have been fixed in commits `a8a43b4` through current working tree.**
-
----
-
-## Critical — FIXED
-
-### BUG-014: `removeLiveAgentHandle` on normal completion destroys resume capability
-
-**File:** `src/runtime/live-session-runtime.ts` (finally block, line ~610)  
-**Commit:** `7a25644` → **Fixed in a8a43b4**
-
-The finally block calls `removeLiveAgentHandle(agentId)` for non-aborted completions. This **deletes the live-agent handle from the registry entirely**, including its terminal status.
-
-**Impact:** After a live-session task completes normally, `steerLiveAgent`, `followUpLiveAgent`, and `resumeLiveAgent` all fail with "Live agent '...' is not registered in this process."
-
-**Fix:** Added `disposeLiveAgentSession()` in `live-agent-manager.ts` — disposes session resources but keeps the handle in the registry for resume/follow-up.
-
----
-
-## High — FIXED
-
-### BUG-015: `withAgentsLock` crashes on `EISDIR` from corrupted lock path
-
-**File:** `src/runtime/crew-agent-records.ts`  
-**Commit:** `c7bd455` → **Fixed in a8a43b4**
-
-`withAgentsLock` only caught `EEXIST`; if lock path was a directory, `EISDIR` crashed the process.
-
-**Fix:** Handle `EISDIR` by removing the directory and retrying.
-
----
-
-## Medium — FIXED
-
-### BUG-016: `markActiveTasksAndAgentsFailed` lacks run-level lock, races with crash recovery
-
-**File:** `src/extension/async-notifier.ts`  
-**Commit:** `d6d466d` → **Fixed in current working tree**
-
-`markActiveTasksAndAgentsFailed` called `saveRunTasks` and `saveCrewAgents` without `withRunLockSync`. Crash recovery (`cancelOrphanedRuns`) uses `withRunLockSync`. Concurrent execution on the same run risks lost updates.
-
-**Fix:** Wrapped `markDeadAsyncRunIfNeeded` mutations inside `withRunLockSync(run, () => { ... })`. Also reloaded fresh manifest inside the lock to avoid operating on stale data.
-
----
-
-## Low — FIXED
-
-### BUG-017: `safeDisposeLiveSession` catches all errors silently
-
-**File:** `src/runtime/live-agent-manager.ts`  
-**Commit:** `7a25644` → **Fixed in current working tree**
-
-Any exception from `dispose()` was swallowed without logging.
-
-**Fix:** Log disposal errors via `logInternalError`.
-
-### BUG-018: `effectiveRuntime` in `run.ts` is manually constructed
-
-**File:** `src/extension/team-tool/run.ts`  
-**Commit:** `2486051` → **Fixed in current working tree**
-
-The async fallback runtime was manually built field-by-field. If `CrewRuntimeCapabilities` gains a new required field, this site won't inherit it.
-
-**Fix:** Use spread of original runtime with overrides: `{ ...runtime, kind: "child-process", steer: false, resume: false, liveToolActivity: false, fallback: "child-process", reason: "..." }`.
-
-### NIT-007: `removeStaleAgentsLock` reads unlimited file size
-
-**File:** `src/runtime/crew-agent-records.ts`  
-**Commit:** `c7bd455` → **Fixed in current working tree**
-
-`fs.readFileSync(lockPath, "utf-8")` reads entire file into memory. Corrupted multi-megabyte file causes memory pressure.
-
-**Fix:** Check `fs.statSync(lockPath).size > 1024` before reading; skip stale removal if oversized.
-
-### NIT-008: `maxCollectedJsonEvents` cap is hardcoded at 200
-
-**File:** `src/runtime/live-session-runtime.ts`  
-**Commit:** `c7bd455` → **Fixed in current working tree**
-
-Long-running agents drop older JSON events needed for debugging/yield detection.
-
-**Fix:** Increased cap from 200 to 1000.
-
----
-
-## Verification
-
-- `npm run typecheck` passes.
-- Targeted tests (`async-notifier`, `isolation-policy`, `live-agent-manager`, `live-session-runtime`, `team-tool-dispatch`) pass: **31 pass / 0 fail**.
-- Full unit test suite not run due to time constraints; no new failures observed in targeted runs.
-
-## Files Changed in Fixes
-
-- `src/extension/async-notifier.ts` — BUG-016 (run lock)
-- `src/runtime/live-agent-manager.ts` — BUG-014 (disposeLiveAgentSession), BUG-017 (log dispose errors)
-- `src/runtime/live-session-runtime.ts` — BUG-014 (use disposeLiveAgentSession), NIT-008 (cap 1000)
-- `src/runtime/crew-agent-records.ts` — BUG-015 (EISDIR), NIT-007 (size cap)
-- `src/extension/team-tool/run.ts` — BUG-018 (spread runtime)
-- `test/unit/live-agent-manager.test.ts` — test for disposeLiveAgentSession