pan-wizard 3.5.2 → 3.7.10

package/commands/pan/experiment.md

@@ -0,0 +1,219 @@
+---
+name: experiment
+group: Self-Improvement
+description: Manage external experiments — scaffold, run, harvest, promote findings back to PAN
+allowed-tools:
+  - Read
+  - Write
+  - Edit
+  - Bash
+  - Grep
+  - Glob
+  - Agent
+---
+
+# /pan:experiment — Cross-Project Self-Improvement Loop
+
+> **Self-protection:** This command **scaffolds external project folders OUTSIDE the PAN source repo** to drive autonomous AI coding sessions against fresh ideas, then harvests the resulting telemetry back into `pan-wizard-core/learnings/`. It is a **PAN-development tool**, not a feature for end-users of PAN to invoke on their own projects.
+
+**Spec:** `docs/specs/self_improvement_loop_featureai.md`
+**ADR:** ADR-0026 (pending W4)
+**Status:** v3.7.0 W1+W2+W3 — scaffolding (`new`/`list`/`manifest`) + external runner (`run`/`status`/`stop`) + harvest (`harvest`/`prune`); W4 adds promote integration with `/pan:learn`.
+
+---
+
+## When to use this
+
+- You have an idea for a small project. You want to know **how PAN drives that build** — what shortcuts the AI takes, where verification fails, which guardrails fire.
+- You want the resulting telemetry to **flow back into PAN's shipped artifacts** so the next release is smarter.
+- You're testing a behavioral change (new `references/` doc, new workflow rule) by running it against a real, isolated build.
+
+## When NOT to use this
+
+- Building production user features. Use `/pan:new-project` and `/pan:exec-phase` directly.
+- Validating a single-file change. The experiment loop is heavy — use `npm test` and `/pan:check`.
+- Inside the PAN source repo. The command **refuses** to scaffold experiments inside `d:\PanWizard\` (or wherever the source is cloned). The experiment root defaults to `~/pan-experiments/`.
+
+---
+
+## Subcommands (W1 shipped)
+
+### `/pan:experiment new <slug> --idea <path>`
+
+Scaffold a new experiment folder. Creates:
+
+```
+<root>/<slug>/
+├── .planning/
+│   ├── idea.md              ← copied from --idea path
+│   └── experiment.json      ← manifest (slug, runtime, created_at, etc.)
+└── .claude/ (or .codex/, .gemini/, .opencode/, .github/)   ← PAN install for chosen runtime
+```
+
+**Flags:**
+
+| Flag | Default | Purpose |
+|------|---------|---------|
+| `--idea <path>` | required | Path to the idea.md doc; copied into the experiment |
+| `--runtime <r>` | `claude` | Which AI coding runtime to install: claude / codex / gemini / opencode / copilot |
+| `--root <path>` | `~/pan-experiments/` | Override the experiment root directory |
+| `--budget <pts>` | `80` | Optional budget cap (saved to manifest, enforced by W2 runner) |
+| `--skip-installer` | `false` | Don't run the PAN installer (dev-only) |
+
+**Slug rules:** lowercase letters, digits, hyphens. Max 40 chars. No leading/trailing hyphen.
+
+**Returns:** JSON with `experiment_id`, `path`, `runtime`, `idea_path`, `created_at`. On failure, returns `{ error: "..." }`.
+
+### `/pan:experiment list`
+
+Enumerate all experiments under the root. Returns `{ experiments: [...], count }` sorted newest-first.
+
+**Flags:**
+
+| Flag | Default | Purpose |
+|------|---------|---------|
+| `--root <path>` | `~/pan-experiments/` | Override the root |
+| `--raw` | `false` | Human-readable output instead of JSON |
+
+### `/pan:experiment manifest <slug>`
+
+Read the manifest for a single experiment. Returns the JSON shape written by `new`.
+
+**Flags:**
+
+| Flag | Default | Purpose |
+|------|---------|---------|
+| `--root <path>` | `~/pan-experiments/` | Override the root |
+
+---
+
+## Subcommands (W2 shipped)
+
+### `/pan:experiment run <slug>`
+
+Spawn the external AI runtime against the experiment folder. **Synchronous** — blocks until the external session exits, hits the timeout, or is stopped.
+
+**Flags:**
+
+| Flag | Default | Purpose |
+|------|---------|---------|
+| `--timeout <sec>` | `1800` (30 min) | Hard timeout in seconds; runner sends SIGTERM at deadline |
+| `--prompt <text>` | `/pan:new-project --auto @.planning/idea.md` | Prompt passed to the external runtime |
+| `--root <path>` | `~/pan-experiments/` | Override the experiment root |
+
+**Returns:** `{ status: "done"\|"failed", stop_reason: "success"\|"error"\|"timeout"\|"manual", exit_code, elapsed_ms, started_at, ended_at }`. Run-state is also persisted to `<experiment>/.planning/run-state.json`.
+
+**Runtime support:** claude / codex / gemini / opencode (via `RUNTIME_RUNNERS` adapter map in `runner.cjs`). GitHub Copilot CLI is **unsupported** for the `run` subcommand — no documented headless prompt mode. Copilot users can still scaffold and harvest manually.
+
+### `/pan:experiment status <slug>`
+
+Read the current `run-state.json` snapshot. Returns the full state object (`status`, `stop_reason`, `exit_code`, `elapsed_ms`, `events`).
+
+### `/pan:experiment stop <slug>`
+
+Gracefully halt a running experiment. Reads pid from `run-state.json`, sends SIGTERM, writes `status: failed, stop_reason: manual` to the run state. Returns the updated state.
+
+If the experiment has already finished, returns the existing run state without error.
+
+## Subcommands (W3 shipped)
+
+### `/pan:experiment harvest <slug>`
+
+Copy the experiment's telemetry into `<source-repo>/experiments/<slug>/` so it can be analyzed and promoted into shipped artifacts.
+
+**What gets harvested** (skipped silently if absent):
+
+- `.planning/idea.md` — the original idea doc
+- `.planning/experiment.json` — scaffold manifest
+- `.planning/state.md` — final project state
+- `.planning/run-state.json` — runner result (status, exit_code, elapsed_ms)
+- `.planning/agent-history.json` — every agent spawn during the build
+- `.planning/optimization/` — full trace data (sessions, reports)
+- `.planning/phases/` — phases the external session created
+
+A `harvest.json` manifest is written at the destination capturing source path, timestamp, total bytes, and the list of harvested paths.
+
+**Flags:**
+
+| Flag | Default | Purpose |
+|------|---------|---------|
+| `--root <path>` | `~/pan-experiments/` | Override the experiment root |
+| `--source-root <path>` | PAN source repo | Override harvest destination |
+| `--force` | `false` | Overwrite an existing harvest at the destination |
+
+**Returns:** `{ experiment_id, harvest_path, harvested_paths: [...], total_bytes, harvested_at, pan_version }`. On conflict without `--force`, returns `{ error }`.
+
+### `/pan:experiment prune <slug>`
+
+Remove the experiment folder after harvest.
+
+**Modes:**
+
+- **Soft** (default): rename to `<root>/<slug>-archived-<ISO-timestamp>` so the data is retained but the slug is freed for reuse
+- **Hard** (`--hard` flag): permanently delete
+
+**Flags:**
+
+| Flag | Default | Purpose |
+|------|---------|---------|
+| `--root <path>` | `~/pan-experiments/` | Override the experiment root |
+| `--hard` | `false` | Permanent deletion (irreversible) |
+
+**Returns:** `{ pruned: <slug>, mode: "soft"|"hard", archive_path? }`.
+
+## Subcommands (W4 — coming soon)
+
+| Subcommand | Wave | Purpose |
+|------------|------|---------|
+| `archive <slug>` | W4 | Alias for `prune` (kept for clarity in scripts) |
+| `delete <slug> --confirm` | W4 | Alias for `prune --hard` with confirmation prompt |
+
+W4 also adds:
+- `/pan:learn --experiment <slug>` — runs pan-optimizer over harvested data
+- `pan-tools learn promote --pattern <id> --scope universal --topic <name>` — extracts a finding into `pan-wizard-core/learnings/{universal,internal}/<topic>.md`
+- `pan-tools learn unpromote/list-promoted` — rollback and inventory
+
+---
+
+## CLI mapping (`pan-tools experiment <sub>`)
+
+```bash
+pan-tools experiment new <slug> --idea <path> [--runtime r] [--root path] [--budget n]
+pan-tools experiment list [--root path] [--raw]
+pan-tools experiment manifest <slug> [--root path]
+```
+
+Examples:
+
+```bash
+# Scaffold a new experiment
+echo "# Idea: Build a markdown linter CLI" > my-idea.md
+# (fill in problem, success, scope, constraints — see template at pan-wizard-core/templates/idea.md)
+pan-tools experiment new md-lint --idea my-idea.md --runtime claude --budget 60
+
+# List all experiments
+pan-tools experiment list
+
+# Inspect one
+pan-tools experiment manifest md-lint
+```
+
+---
+
+## Safety guards
+
+- **Never inside source repo.** `newExperiment` refuses to write to `d:\PanWizard\` (or wherever the PAN source is). Mirrors `bin/install.js` `PAN_SOURCE_ROOT` guard.
+- **No clobber.** Refuses to scaffold over an existing experiment folder of the same slug.
+- **Slug validation.** Lowercase + digits + hyphens, max 40 chars. Rejects uppercase, spaces, special characters.
+- **Idea path validation.** Errors if the `--idea` file doesn't exist.
+
+## Related
+
+- `pan-wizard-core/templates/idea.md` — idea doc template
+- `pan-wizard-core/bin/lib/experiment.cjs` — implementation
+- `pan-wizard-core/learnings/README.md` — where harvested findings go
+- `docs/specs/self_improvement_loop_featureai.md` — full design
+
+## Runtime support
+
+Works in all 5 runtimes (Claude / Codex / Gemini / OpenCode / Copilot). The W2 external runner (subprocess invocation of the external session) supports Claude / Codex / Gemini / OpenCode; GitHub Copilot CLI lacks a headless prompt mode and is opt-out for the `run` subcommand.

package/commands/pan/health.md

@@ -2,7 +2,7 @@
 name: pan:health
 group: System
 description: Diagnose planning directory health and optionally repair issues
-argument-hint: [--repair]
+argument-hint: "[--repair]"
 allowed-tools:
   - Read
   - Bash

package/commands/pan/learn.md

@@ -1,3 +1,15 @@
+---
+name: pan:learn
+group: Self-Improvement
+description: Analyze trace sessions or harvested experiments via pan-optimizer; generate ranked optimization reports
+allowed-tools:
+  - Read
+  - Bash
+  - Glob
+  - Grep
+  - Task
+---
+
 # /pan:learn
 
 Analyze the most recent trace session and generate an optimization report.
@@ -6,11 +18,13 @@ Analyze the most recent trace session and generate an optimization report.
 ```
 /pan:learn
 /pan:learn --session <session-id>
+/pan:learn --experiment <slug>
 /pan:learn --apply
 ```
 
 **Flags:**
 - `--session <id>` — analyze a specific session instead of the most recent
+- `--experiment <slug>` *(v3.7.0+, W3)* — analyze a harvested experiment instead of the current project's traces. Reads from `<source-repo>/experiments/<slug>/.planning/optimization/` and writes the report to `<source-repo>/experiments/<slug>/learnings/report-<timestamp>.md`. Used by the self-improvement loop. Run `/pan:experiment harvest <slug>` first.
 - `--apply` — automatically apply safe optimizations after generating the report (equivalent to running `/pan:optimize apply` immediately after)
 
 **What it does:**
@@ -56,6 +70,6 @@ The optimization report in `.planning/optimization/reports/` contains:
 → Needs review: 2 prompt improvements, 1 workflow gap
 ```
 
-**See also:** `/pan:optimize`, `/pan:exec-phase`
+**See also:** `/pan:optimize`, `/pan:exec-phase`, `/pan:experiment` (v3.7.0+ self-improvement loop)
 
 Follow the workflow at `.claude/workflows/learn.md` (or `pan-wizard-core/workflows/learn.md`).

package/commands/pan/optimize.md

@@ -1,3 +1,16 @@
+---
+name: pan:optimize
+group: Self-Improvement
+description: Manage the circular optimization loop — apply recommendations, view stats, list reports, manage trace sessions
+allowed-tools:
+  - Read
+  - Write
+  - Edit
+  - Bash
+  - Glob
+  - Grep
+---
+
 # /pan:optimize
 
 Manage the circular optimization loop: apply recommendations, view stats, list reports.

package/commands/pan/patches.md

@@ -1,6 +1,15 @@
 ---
+name: pan:patches
+group: System
 description: Reapply local modifications after a PAN update
-allowed-tools: Read, Write, Edit, Bash, Glob, Grep, AskUserQuestion
+allowed-tools:
+  - Read
+  - Write
+  - Edit
+  - Bash
+  - Glob
+  - Grep
+  - AskUserQuestion
 ---
 
 <purpose>

package/commands/pan/phase-tests.md

@@ -12,10 +12,7 @@ allowed-tools:
   - Grep
   - Task
   - AskUserQuestion
-argument-instructions: |
-  Parse the argument as a phase number (integer, decimal, or letter-suffix), plus optional free-text instructions.
-  Example: /pan:phase-tests 12
-  Example: /pan:phase-tests 12 focus on edge cases in the pricing module
+argument-instructions: "Parse <phase> as a phase number (integer, decimal like 1.1, or letter-suffix like 1a) followed by optional free-text instructions. Examples: /pan:phase-tests 12 — /pan:phase-tests 12 focus on edge cases in the pricing module"
 ---
 <objective>
 Generate unit and E2E tests for a completed phase, using its summary.md, context.md, and verification.md as specifications.

package/commands/pan/todo-add.md

@@ -2,7 +2,7 @@
 name: pan:todo-add
 group: System
 description: Capture idea or task as todo from current conversation context
-argument-hint: [optional description]
+argument-hint: "[optional description]"
 allowed-tools:
   - Read
   - Write

package/commands/pan/todo-check.md

@@ -2,7 +2,7 @@
 name: pan:todo-check
 group: System
 description: List pending todos and select one to work on
-argument-hint: [area filter]
+argument-hint: "[area filter]"
 allowed-tools:
   - Read
   - Write

package/hooks/dist/pan-cost-logger.js

@@ -33,16 +33,32 @@ function buildCostRecord(data, cwd) {
   // Only log actual subagent stops; ignore other Stop variants.
   if (data.hook_event_name && data.hook_event_name !== 'SubagentStop') return null;
 
+  // P-1805 (v3.7.8): if data.usage is missing/empty (Claude Code headless mode
+  // doesn't include it in the SubagentStop payload), fall back to reading the
+  // transcript_path JSONL and summing usage across the subagent's messages.
+  // Same approach as pan-trace-logger.js for consistency.
+  let inputTokens = extractNumber(data.usage, 'input_tokens');
+  let outputTokens = extractNumber(data.usage, 'output_tokens');
+  let cacheRead = extractNumber(data.usage, 'cache_read_input_tokens');
+  let cacheWrite = extractNumber(data.usage, 'cache_creation_input_tokens');
+  if ((inputTokens + outputTokens + cacheRead + cacheWrite) === 0 && data.transcript_path) {
+    const fromTranscript = readUsageFromTranscript(data.transcript_path, data.session_id);
+    inputTokens = fromTranscript.input_tokens;
+    outputTokens = fromTranscript.output_tokens;
+    cacheRead = fromTranscript.cache_read_input_tokens;
+    cacheWrite = fromTranscript.cache_creation_input_tokens;
+  }
+
   const record = {
     ts: new Date().toISOString(),
     agent: data.agent_type || data.subagent_type || null,
     command: null,
     model: data.model || null,
     tier: null,
-    input_tokens: extractNumber(data.usage, 'input_tokens') || 0,
-    output_tokens: extractNumber(data.usage, 'output_tokens') || 0,
-    cache_read_tokens: extractNumber(data.usage, 'cache_read_input_tokens') || 0,
-    cache_write_tokens: extractNumber(data.usage, 'cache_creation_input_tokens') || 0,
+    input_tokens: inputTokens,
+    output_tokens: outputTokens,
+    cache_read_tokens: cacheRead,
+    cache_write_tokens: cacheWrite,
     cost_usd: null,
     phase: data.phase || null,
     session: data.session_id || null,
@@ -58,6 +74,40 @@ function extractNumber(obj, key) {
   return typeof v === 'number' ? v : 0;
 }
 
+/**
+ * P-1805 (v3.7.8): read transcript JSONL and sum usage across all assistant
+ * messages belonging to the subagent's session. Mirrors the helper in
+ * pan-trace-logger.js. Returns zeros if transcript missing/unreadable.
+ */
+function readUsageFromTranscript(transcriptPath, sessionId) {
+  const totals = {
+    input_tokens: 0,
+    output_tokens: 0,
+    cache_read_input_tokens: 0,
+    cache_creation_input_tokens: 0,
+  };
+  if (!transcriptPath || typeof transcriptPath !== 'string') return totals;
+  let raw;
+  try { raw = fs.readFileSync(transcriptPath, 'utf-8'); } catch { return totals; }
+  for (const line of raw.split('\n')) {
+    if (!line) continue;
+    let entry;
+    try { entry = JSON.parse(line); } catch { continue; }
+    if (sessionId && entry.session_id && entry.session_id !== sessionId) continue;
+    const usage = entry.usage
+      || entry.message?.usage
+      || entry.response?.usage
+      || (entry.type === 'assistant' && entry.message?.usage)
+      || null;
+    if (!usage || typeof usage !== 'object') continue;
+    totals.input_tokens += extractNumber(usage, 'input_tokens');
+    totals.output_tokens += extractNumber(usage, 'output_tokens');
+    totals.cache_read_input_tokens += extractNumber(usage, 'cache_read_input_tokens');
+    totals.cache_creation_input_tokens += extractNumber(usage, 'cache_creation_input_tokens');
+  }
+  return totals;
+}
+
 /**
  * Append record to .planning/metrics/tokens.jsonl. Silently succeeds
  * even if the file or directory can't be written — hook must not block.

package/hooks/dist/pan-trace-logger.js

@@ -80,6 +80,65 @@ function extractNumber(obj, key) {
   return typeof v === 'number' ? v : 0;
 }
 
+/**
+ * P-1805 (v3.7.8): extract usage totals by reading the SubagentStop transcript.
+ * The hook payload from Claude Code in headless mode does NOT include
+ * `data.usage` — only `transcript_path`. This function reads the transcript
+ * JSONL file and sums the `usage` fields across all assistant messages
+ * belonging to the just-completed subagent (its session_id is in `data.session_id`).
+ *
+ * Returns the SUM of input/output/cache tokens for the subagent's whole
+ * conversation. Agents that finish in one Claude API call will have one
+ * usage record; agents with multi-turn tool use will have several.
+ *
+ * Returns `{input_tokens, output_tokens, cache_read_input_tokens, cache_creation_input_tokens}`
+ * with all zeros if the transcript is unreadable / missing — same fallback
+ * shape as the original behavior, so callers don't break.
+ *
+ * @param {string} transcriptPath - Absolute path to the transcript JSONL file
+ * @param {string} [sessionId] - Optional subagent session_id to filter on
+ * @returns {Object} usage totals object
+ */
+function readUsageFromTranscript(transcriptPath, sessionId) {
+  const totals = {
+    input_tokens: 0,
+    output_tokens: 0,
+    cache_read_input_tokens: 0,
+    cache_creation_input_tokens: 0,
+  };
+  if (!transcriptPath || typeof transcriptPath !== 'string') return totals;
+  let raw;
+  try {
+    raw = fs.readFileSync(transcriptPath, 'utf-8');
+  } catch {
+    return totals;
+  }
+  for (const line of raw.split('\n')) {
+    if (!line) continue;
+    let entry;
+    try {
+      entry = JSON.parse(line);
+    } catch {
+      continue;
+    }
+    // Filter to entries from this subagent if a session_id is provided.
+    // The transcript may include parent + child traffic; session_id discriminates.
+    if (sessionId && entry.session_id && entry.session_id !== sessionId) continue;
+    // Usage typically lives on assistant message records.
+    const usage = entry.usage
+      || entry.message?.usage
+      || entry.response?.usage
+      || (entry.type === 'assistant' && entry.message?.usage)
+      || null;
+    if (!usage || typeof usage !== 'object') continue;
+    totals.input_tokens += extractNumber(usage, 'input_tokens');
+    totals.output_tokens += extractNumber(usage, 'output_tokens');
+    totals.cache_read_input_tokens += extractNumber(usage, 'cache_read_input_tokens');
+    totals.cache_creation_input_tokens += extractNumber(usage, 'cache_creation_input_tokens');
+  }
+  return totals;
+}
+
 /**
  * Build trace event(s) from a SubagentStop payload.
  * Pure function — no side effects.
@@ -93,9 +152,19 @@ function buildTraceEvents(data, sessionId) {
 
   const ts = new Date().toISOString();
   const agent = data.agent_type || data.subagent_type || 'unknown';
-  const inputTokens = extractNumber(data.usage, 'input_tokens');
-  const outputTokens = extractNumber(data.usage, 'output_tokens');
-  const cacheRead = extractNumber(data.usage, 'cache_read_input_tokens');
+
+  // P-1805: prefer usage from data.usage when present (interactive Claude Code path).
+  // Fall back to reading the transcript file (headless `claude -p` path — usage
+  // not in payload but discoverable via transcript_path).
+  let inputTokens = extractNumber(data.usage, 'input_tokens');
+  let outputTokens = extractNumber(data.usage, 'output_tokens');
+  let cacheRead = extractNumber(data.usage, 'cache_read_input_tokens');
+  if ((inputTokens + outputTokens + cacheRead) === 0 && data.transcript_path) {
+    const fromTranscript = readUsageFromTranscript(data.transcript_path, data.session_id);
+    inputTokens = fromTranscript.input_tokens;
+    outputTokens = fromTranscript.output_tokens;
+    cacheRead = fromTranscript.cache_read_input_tokens;
+  }
   const totalTokens = inputTokens + outputTokens;
 
   const events = [];

package/package.json

@@ -1,66 +1,67 @@
-{
-  "name": "pan-wizard",
-  "version": "3.5.2",
-  "description": "A lightweight workflow automation and context engineering system for Claude Code, OpenCode, Gemini CLI, Codex, and Copilot CLI.",
-  "bin": {
-    "pan-wizard": "bin/install.js"
-  },
-  "files": [
-    "bin",
-    "commands",
-    "pan-wizard-core",
-    "agents",
-    "hooks/dist",
-    "scripts",
-    "assets"
-  ],
-  "keywords": [
-    "claude",
-    "claude-code",
-    "ai",
-    "workflow-automation",
-    "context-engineering",
-    "project-automation",
-    "gemini",
-    "gemini-cli",
-    "codex",
-    "codex-cli",
-    "copilot",
-    "copilot-cli",
-    "github-copilot"
-  ],
-  "author": "PAN Wizard Contributors",
-  "contributors": [
-    {
-      "name": "oharms",
-      "url": "https://github.com/oharms"
-    }
-  ],
-  "license": "MIT",
-  "repository": {
-    "type": "git",
-    "url": "git+https://github.com/oharms/PanWizard.git"
-  },
-  "homepage": "https://github.com/oharms/PanWizard#readme",
-  "bugs": {
-    "url": "https://github.com/oharms/PanWizard/issues"
-  },
-  "engines": {
-    "node": ">=16.7.0"
-  },
-  "devDependencies": {
-    "@playwright/test": "^1.58.2",
-    "@vscode/test-electron": "^2.5.2",
-    "esbuild": "^0.24.0"
-  },
-  "scripts": {
-    "build:hooks": "node scripts/build-hooks.js",
-    "prepublishOnly": "npm run build:hooks",
-    "test": "node --test tests/*.test.cjs",
-    "test:scenarios": "node --test tests/scenarios/*.test.cjs",
-    "test:all": "node --test tests/*.test.cjs tests/scenarios/*.test.cjs",
-    "test:e2e": "node --test tests/scenarios/*.test.cjs",
-    "test:vscode": "npx playwright test --config tests/e2e/playwright.config.mjs",
-    "test:watch": "node --test --watch tests/*.test.cjs"
-  }
-}
+{
+  "name": "pan-wizard",
+  "version": "3.7.10",
+  "description": "A lightweight workflow automation and context engineering system for Claude Code, OpenCode, Gemini CLI, Codex, and Copilot CLI.",
+  "bin": {
+    "pan-wizard": "bin/install.js"
+  },
+  "files": [
+    "bin",
+    "commands",
+    "pan-wizard-core",
+    "agents",
+    "hooks/dist",
+    "scripts",
+    "assets"
+  ],
+  "keywords": [
+    "claude",
+    "claude-code",
+    "ai",
+    "workflow-automation",
+    "context-engineering",
+    "project-automation",
+    "gemini",
+    "gemini-cli",
+    "codex",
+    "codex-cli",
+    "copilot",
+    "copilot-cli",
+    "github-copilot"
+  ],
+  "author": "PAN Wizard Contributors",
+  "contributors": [
+    {
+      "name": "oharms",
+      "url": "https://github.com/oharms"
+    }
+  ],
+  "license": "MIT",
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/oharms/PanWizard.git"
+  },
+  "homepage": "https://github.com/oharms/PanWizard#readme",
+  "bugs": {
+    "url": "https://github.com/oharms/PanWizard/issues"
+  },
+  "engines": {
+    "node": ">=16.7.0"
+  },
+  "devDependencies": {
+    "@playwright/test": "^1.58.2",
+    "@vscode/test-electron": "^2.5.2",
+    "esbuild": "^0.24.0"
+  },
+  "scripts": {
+    "build:hooks": "node scripts/build-hooks.js",
+    "release:check": "node scripts/release-check.js",
+    "prepublishOnly": "node scripts/release-check.js",
+    "test": "node --test tests/*.test.cjs",
+    "test:scenarios": "node --test tests/scenarios/*.test.cjs",
+    "test:all": "node --test tests/*.test.cjs tests/scenarios/*.test.cjs",
+    "test:e2e": "node --test tests/scenarios/*.test.cjs",
+    "test:vscode": "npx playwright test --config tests/e2e/playwright.config.mjs",
+    "test:watch": "node --test --watch tests/*.test.cjs"
+  }
+}

package/pan-wizard-core/bin/lib/commands.cjs

@@ -379,6 +379,11 @@ function runCommitSafetyChecks(cwd, config, force) {
 function cmdCommit(cwd, message, files, raw, amend, opts) {
   const commitType = opts && opts.type;
   const force = opts && opts.force;
+  // P-EXP-001 follow-up (v3.7.10): when failOnError is true, commit_failed
+  // exits non-zero so callers (especially autonomous loops) detect the silent-
+  // failure case where git refused (e.g. missing identity) and the loop would
+  // otherwise keep going thinking the artifact landed.
+  const failOnError = opts && opts.failOnError;
 
   if (!isGitRepo(cwd)) {
     output({ committed: false, hash: null, reason: 'not_a_git_repo', hint: 'Run git init to initialize a repository' }, raw, 'not a git repo');
@@ -425,6 +430,9 @@ function cmdCommit(cwd, message, files, raw, amend, opts) {
       output({ committed: false, hash: null, reason: 'nothing_to_commit' }, raw, 'nothing');
       return;
     }
+    if (failOnError) {
+      error('commit_failed: ' + (commitResult.stderr || 'unknown git error').trim());
+    }
     output({ committed: false, hash: null, reason: 'commit_failed', error: commitResult.stderr }, raw, 'failed');
     return;
   }