gm-skill 2.0.1148 → 2.0.1150

package/README.md

@@ -35,7 +35,7 @@ An earlier generation fanned out fifteen per-platform downstream repos (gm-cc, g
 
 ## Version
 
-`2.0.1148` — auto-bumped from the canonical `gm` repo. Every push to `AnEntrypoint/gm` (or any cascading sibling crate) republishes this package.
+`2.0.1150` — auto-bumped from the canonical `gm` repo. Every push to `AnEntrypoint/gm` (or any cascading sibling crate) republishes this package.
 
 ## Source of truth
 

package/bin/plugkit.version

@@ -1 +1 @@
-0.1.413
+0.1.415

package/bin/plugkit.wasm.sha256

@@ -1 +1 @@
-4032ac9258c6fb79106ef9e02a40c9cba7a509a6e0507d4d0c062645b25799a6  plugkit.wasm
+3af3d80f571bec77e28471184811226f865a150cfcbecef00b6d90e2ed90c140  plugkit.wasm

package/gm-plugkit/plugkit-wasm-wrapper.js

@@ -14,6 +14,73 @@ const __dirname = path.dirname(__filename);
 const KV_DIR = path.join(os.homedir(), '.claude', 'gm-tools', 'kv');
 fs.mkdirSync(KV_DIR, { recursive: true });
 
+const GM_LOG_ROOT = process.env.GM_LOG_DIR || path.join(os.homedir(), '.claude', 'gm-log');
+const ORCHESTRATOR_VERBS = new Set(['instruction', 'transition', 'phase-status', 'prd-add', 'prd-resolve', 'prd-list', 'mutable-add', 'mutable-resolve', 'mutable-list', 'memorize-fire', 'residual-scan', 'auto-recall']);
+
+function logEvent(sub, event, fields) {
+  if (process.env.GM_LOG_DISABLE) return;
+  try {
+    const day = new Date().toISOString().slice(0, 10);
+    const dir = path.join(GM_LOG_ROOT, day);
+    fs.mkdirSync(dir, { recursive: true });
+    const line = JSON.stringify({
+      ts: new Date().toISOString(),
+      sub,
+      event,
+      pid: process.pid,
+      sess: process.env.CLAUDE_SESSION_ID || process.env.GM_SESSION_ID || '',
+      ...fields,
+    });
+    fs.appendFileSync(path.join(dir, `${sub}.jsonl`), line + '\n');
+  } catch (_) {}
+}
+
+function emitOrchestratorEvents(verb, taskBase, resultStr) {
+  if (!ORCHESTRATOR_VERBS.has(verb)) return;
+  let parsed;
+  try { parsed = JSON.parse(resultStr); } catch (_) { return; }
+  if (!parsed || parsed.ok !== true) {
+    logEvent('plugkit', 'orchestrator.error', { verb, task: taskBase, error: parsed && parsed.error ? String(parsed.error) : 'unknown' });
+    return;
+  }
+  const data = parsed.data || {};
+  switch (verb) {
+    case 'transition':
+      logEvent('plugkit', 'phase.transitioned', { task: taskBase, phase: data.phase, next_skill: data.nextSkill, recall_count: Array.isArray(data.recall_hits) ? data.recall_hits.length : 0 });
+      break;
+    case 'instruction':
+      logEvent('plugkit', 'instruction.served', { task: taskBase, phase: data.phase, prd_pending: data.prd_pending_count, mutables_pending: Array.isArray(data.mutables_pending) ? data.mutables_pending.length : 0, next_phase_hint: data.next_phase_hint });
+      break;
+    case 'phase-status':
+      logEvent('plugkit', 'phase.status', { task: taskBase, phase: data.phase, last_skill: data.last_skill });
+      break;
+    case 'prd-add':
+      logEvent('plugkit', 'prd.added', { task: taskBase, id: data.added });
+      break;
+    case 'prd-resolve':
+      logEvent('plugkit', 'prd.resolved', { task: taskBase, id: data.resolved });
+      break;
+    case 'mutable-add':
+      logEvent('plugkit', 'mutable.added', { task: taskBase, id: data.added });
+      break;
+    case 'mutable-resolve':
+      logEvent('plugkit', 'mutable.resolved', { task: taskBase, id: data.resolved, memorize_spool: data.memorize_spool });
+      break;
+    case 'memorize-fire':
+      logEvent('plugkit', 'memorize.fired', { task: taskBase, key: data.key, namespace: data.namespace, bytes: data.bytes });
+      break;
+    case 'residual-scan':
+      if (data.scan === 'fired') logEvent('plugkit', 'residual.fired', { task: taskBase, marker: data.marker });
+      else logEvent('plugkit', 'residual.skipped', { task: taskBase, reason: data.reason });
+      break;
+    case 'auto-recall':
+      logEvent('plugkit', 'auto_recall.hits', { task: taskBase, count: Array.isArray(data.hits) ? data.hits.length : 0 });
+      break;
+    default:
+      break;
+  }
+}
+
 const TMP_DIR = os.tmpdir();
 const BROWSER_PORTS_FILE = path.join(TMP_DIR, 'plugkit-browser-ports.json');
 const BROWSER_SESSIONS_FILE = path.join(TMP_DIR, 'plugkit-browser-sessions.json');
@@ -656,8 +723,10 @@ async function runSpoolWatcher(instance, spoolDir) {
     }
   } catch (e) { console.error(`[plugkit-wasm] wrapper self-install failed: ${e.message}`); }
 
-  console.log(`[plugkit-wasm] plugkit v${resolveVersion(instance)} (wasm)`);
+  const _bootVersion = resolveVersion(instance);
+  console.log(`[plugkit-wasm] plugkit v${_bootVersion} (wasm)`);
   console.log(`[plugkit-wasm] watching ${inDir}`);
+  logEvent('plugkit', 'watcher.boot', { version: _bootVersion, in_dir: inDir, out_dir: outDir, spool_dir: spoolDir });
 
   const PROCESSED_MAX = 10000;
   const processed = new Map();
@@ -692,6 +761,7 @@ async function runSpoolWatcher(instance, spoolDir) {
 
       const t0 = Date.now();
       console.log(`[dispatch] → verb=${verb} task=${taskBase} body=${bodyBytes.length}b`);
+      logEvent('plugkit', 'dispatch.start', { verb, task: taskBase, body_bytes: bodyBytes.length, cwd: process.cwd() });
 
       const verbPtr = instance.exports.plugkit_alloc(verbBytes.length);
       const bodyPtr = instance.exports.plugkit_alloc(bodyBytes.length);
@@ -707,7 +777,10 @@ async function runSpoolWatcher(instance, spoolDir) {
 
       const outName = dir === '.' ? `${taskBase}.json` : `${verb}-${taskBase}.json`;
       fs.writeFileSync(path.join(outDir, outName), resultStr);
-      console.log(`[dispatch] ← verb=${verb} task=${taskBase} ms=${Date.now() - t0} out=${resultStr.length}b`);
+      const dur_ms = Date.now() - t0;
+      console.log(`[dispatch] ← verb=${verb} task=${taskBase} ms=${dur_ms} out=${resultStr.length}b`);
+      logEvent('plugkit', 'dispatch.end', { verb, task: taskBase, dur_ms, out_bytes: resultStr.length });
+      emitOrchestratorEvents(verb, taskBase, resultStr);
 
       try { instance.exports.plugkit_free(verbPtr, verbBytes.length); } catch (_) {}
       try { instance.exports.plugkit_free(bodyPtr, bodyBytes.length); } catch (_) {}
@@ -727,6 +800,7 @@ async function runSpoolWatcher(instance, spoolDir) {
       } catch (_) {}
       try { fs.unlinkSync(filePath); } catch (_) {}
       unmarkProcessed(key);
+      logEvent('plugkit', 'dispatch.error', { verb, task: taskBase, error: String(e && e.message || e) });
     }
   };
 
@@ -819,8 +893,14 @@ async function runSpoolWatcher(instance, spoolDir) {
           if (s.mtimeMs < cutoff) { fs.unlinkSync(fp); swept++; }
         } catch (e) { console.error(`[retention] failed to sweep ${entry}: ${e.message}`); }
       }
-      if (swept > 0) console.log(`[retention] swept ${swept} out/ files older than 1h`);
-    } catch (e) { console.error(`[retention] sweep error: ${e.message}`); }
+      if (swept > 0) {
+        console.log(`[retention] swept ${swept} out/ files older than 1h`);
+        logEvent('plugkit', 'sweep.retention', { swept });
+      }
+    } catch (e) {
+      console.error(`[retention] sweep error: ${e.message}`);
+      logEvent('plugkit', 'sweep.retention.error', { error: String(e.message || e) });
+    }
   }, 60_000);
 
   setInterval(() => {
@@ -848,8 +928,14 @@ async function runSpoolWatcher(instance, spoolDir) {
         }
       };
       walk(inDir);
-      if (stale > 0) console.log(`[stale-sweep] failed ${stale} orphaned inputs`);
-    } catch (e) { console.error(`[stale-sweep] sweep error: ${e.message}`); }
+      if (stale > 0) {
+        console.log(`[stale-sweep] failed ${stale} orphaned inputs`);
+        logEvent('plugkit', 'sweep.stale', { stale });
+      }
+    } catch (e) {
+      console.error(`[stale-sweep] sweep error: ${e.message}`);
+      logEvent('plugkit', 'sweep.stale.error', { error: String(e.message || e) });
+    }
   }, 300_000);
 
   const existing = walkDir(inDir);

package/gm.json

@@ -1,6 +1,6 @@
 {
   "name": "gm",
-  "version": "2.0.1148",
+  "version": "2.0.1150",
   "description": "Spool-dispatch orchestration engine with unified state machine, skills, and automated git enforcement",
   "author": "AnEntrypoint",
   "license": "MIT",
@@ -17,5 +17,5 @@
   "publishConfig": {
     "access": "public"
   },
-  "plugkitVersion": "0.1.413"
+  "plugkitVersion": "0.1.415"
 }

package/lib/spool-dispatch.js

@@ -1,6 +1,54 @@
 const fs = require('fs');
 const path = require('path');
 const os = require('os');
+const { spawnSync } = require('child_process');
+
+const GM_LOG_ROOT = process.env.GM_LOG_DIR || path.join(os.homedir(), '.claude', 'gm-log');
+
+function logDeviation(event, fields) {
+  if (process.env.GM_LOG_DISABLE) return;
+  try {
+    const day = new Date().toISOString().slice(0, 10);
+    const dir = path.join(GM_LOG_ROOT, day);
+    fs.mkdirSync(dir, { recursive: true });
+    const line = JSON.stringify({
+      ts: new Date().toISOString(),
+      sub: 'hook',
+      event,
+      pid: process.pid,
+      sess: process.env.CLAUDE_SESSION_ID || process.env.GM_SESSION_ID || '',
+      cwd: process.cwd(),
+      ...fields,
+    });
+    fs.appendFileSync(path.join(dir, 'hook.jsonl'), line + '\n');
+  } catch (_) {}
+}
+
+function isWorktreeDirty(cwd) {
+  try {
+    const r = spawnSync('git', ['status', '--porcelain'], {
+      cwd: cwd || process.cwd(), encoding: 'utf8', timeout: 1500, windowsHide: true
+    });
+    if (r.status !== 0) return { dirty: false, files: [], available: false };
+    const lines = r.stdout.split('\n').filter(l => l.length > 0);
+    return { dirty: lines.length > 0, files: lines, available: true };
+  } catch (_) {
+    return { dirty: false, files: [], available: false };
+  }
+}
+
+function hasUnpushedCommits(cwd) {
+  try {
+    const r = spawnSync('git', ['log', '@{u}..HEAD', '--oneline'], {
+      cwd: cwd || process.cwd(), encoding: 'utf8', timeout: 1500, windowsHide: true
+    });
+    if (r.status !== 0) return { unpushed: false, count: 0, available: false };
+    const lines = r.stdout.split('\n').filter(l => l.length > 0);
+    return { unpushed: lines.length > 0, count: lines.length, available: true };
+  } catch (_) {
+    return { unpushed: false, count: 0, available: false };
+  }
+}
 
 async function dispatchSpool(cmd, lang, body, timeoutMs, sessionId) {
   const taskId = `${Date.now()}-${Math.random().toString(16).slice(2, 8)}`;
@@ -72,16 +120,82 @@ async function pollForCompletion(jsonFile, timeoutMs, taskId) {
   };
 }
 
-function checkDispatchGates(sessionId, operation) {
-  const gm = path.join(process.cwd(), '.gm');
+function sessionMarkerPath(sessionId, kind) {
+  const cwd = process.cwd();
+  return path.join(cwd, '.gm', 'exec-spool', `.session-${kind}-${sessionId || 'anon'}`);
+}
+
+function hasDispatchedInstruction(sessionId) {
+  try {
+    const outDir = path.join(process.cwd(), '.gm', 'exec-spool', 'out');
+    if (!fs.existsSync(outDir)) return false;
+    for (const f of fs.readdirSync(outDir)) {
+      if (f.startsWith('instruction-')) return true;
+    }
+  } catch (_) {}
+  return fs.existsSync(sessionMarkerPath(sessionId, 'instruction-seen'));
+}
+
+function markInstructionSeen(sessionId) {
+  try {
+    fs.mkdirSync(path.dirname(sessionMarkerPath(sessionId, 'instruction-seen')), { recursive: true });
+    fs.writeFileSync(sessionMarkerPath(sessionId, 'instruction-seen'), String(Date.now()));
+  } catch (_) {}
+}
+
+function checkDispatchGates(sessionId, operation, extra) {
+  const cwd = process.cwd();
+  const gm = path.join(cwd, '.gm');
   const prdPath = path.join(gm, 'prd.yml');
   const mutsPath = path.join(gm, 'mutables.yml');
   const needsGmPath = path.join(gm, 'needs-gm');
   const gmFiredPath = path.join(gm, `gm-fired-${sessionId}`);
 
+  if (['stop', 'complete'].includes(operation)) {
+    const residuals = [];
+    if (fs.existsSync(prdPath)) {
+      try {
+        const content = fs.readFileSync(prdPath, 'utf8');
+        if (content.includes('status: pending') || content.includes('status: in_progress')) {
+          residuals.push('PRD has open items — resolve or name-and-stop before declaring done');
+        }
+      } catch (_) {}
+    }
+    if (fs.existsSync(mutsPath)) {
+      try {
+        const content = fs.readFileSync(mutsPath, 'utf8');
+        if (content.includes('status: unknown')) {
+          residuals.push('unresolved mutables present — resolve with witness_evidence before declaring done');
+        }
+      } catch (_) {}
+    }
+    const dirty = isWorktreeDirty(cwd);
+    if (dirty.available && dirty.dirty) {
+      residuals.push(`worktree dirty (${dirty.files.length} file${dirty.files.length === 1 ? '' : 's'}) — commit and push before declaring done`);
+    }
+    const unpushed = hasUnpushedCommits(cwd);
+    if (unpushed.available && unpushed.unpushed) {
+      residuals.push(`${unpushed.count} unpushed commit${unpushed.count === 1 ? '' : 's'} — push to remote before declaring done`);
+    }
+    if (residuals.length > 0) {
+      logDeviation('deviation.gate-deny', { operation, reason: 'stop-gate residuals', residuals });
+      return { allowed: false, reason: `stop-gate residuals: ${residuals.join('; ')}`, residuals };
+    }
+    return { allowed: true };
+  }
+
+  if (['write', 'edit'].includes(operation) && !hasDispatchedInstruction(sessionId)) {
+    logDeviation('deviation.write-before-instruction', { operation, sessionId });
+  }
+
+  if (operation === 'mutable-resolve' && extra && (!extra.witness_evidence || String(extra.witness_evidence).trim() === '')) {
+    logDeviation('deviation.mutable-without-evidence', { mutable_id: extra.id || null });
+  }
+
   if (!['write', 'edit', 'git'].includes(operation)) return { allowed: true };
 
   if (fs.existsSync(prdPath) && fs.existsSync(needsGmPath) && !fs.existsSync(gmFiredPath)) {
+    logDeviation('deviation.gate-deny', { operation, reason: 'gm orchestration in progress' });
     return { allowed: false, reason: 'gm orchestration in progress; skills must complete work before tools execute' };
   }
 
@@ -89,6 +203,7 @@ function checkDispatchGates(sessionId, operation) {
     try {
       const content = fs.readFileSync(mutsPath, 'utf8');
       if (content.includes('status: unknown')) {
+        logDeviation('deviation.gate-deny', { operation, reason: 'unresolved mutables' });
         return { allowed: false, reason: 'unresolved mutables block tool execution; resolve all mutables before proceeding' };
       }
     } catch (_) {}
@@ -97,4 +212,4 @@ function checkDispatchGates(sessionId, operation) {
   return { allowed: true };
 }
 
-module.exports = { dispatchSpool, checkDispatchGates };
+module.exports = { dispatchSpool, checkDispatchGates, isWorktreeDirty, hasUnpushedCommits, logDeviation, markInstructionSeen, hasDispatchedInstruction };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "gm-skill",
-  "version": "2.0.1148",
+  "version": "2.0.1150",
   "description": "Canonical universal harness — AI-native software engineering via skill-driven orchestration; bootstraps plugkit for task execution and session isolation. Install in any AI coding agent host.",
   "author": "AnEntrypoint",
   "license": "MIT",
@@ -39,7 +39,7 @@
     "gm.json"
   ],
   "dependencies": {
-    "gm-plugkit": "^2.0.1148"
+    "gm-plugkit": "^2.0.1150"
   },
   "engines": {
     "node": ">=16.0.0"

package/skills/gm-skill/SKILL.md

@@ -6,6 +6,14 @@ allowed-tools: Skill, Read, Write, Bash(node *), Bash(bun *)
 
 # gm — single entry point
 
+Every turn begins with `instruction`. Write `.gm/exec-spool/in/instruction/<N>.txt` (empty body) and read the response — its `instruction` field is the active phase prose, its `prd_items` is the live work plan, its `mutables_pending` is what blocks the EMIT gate, its `recall_hits` is the prior memory already loaded for you. Follow that prose imperatively until the phase's exit condition is met, then `transition`. The user's request is the authorization; the PRD is the receipt; the chain runs PLAN → EXECUTE → EMIT → VERIFY → COMPLETE without re-asking between phases.
+
+Direct Read/Edit/Bash before the first `instruction` dispatch is the failure mode this skill prevents. If you find yourself reaching for a tool to "just take a quick look", you have already dropped the loop — dispatch `instruction` first, then act on its prose.
+
+Fix on Sight. Every unexpected output, broken neighbor, lockfile drift, suppressed error, or stale generated file surfaced during work becomes a PRD item the same turn and is executed before COMPLETE. Surprise is a new mutable, not noise to absorb. Stubs, mocks, placeholder returns, fixture-only paths, and "always succeeds" code paths are forbidden in shipped material; acceptance is real input through real code into real output, witnessed. Edits to code that runs in a browser require a live `browser` verb witness in the same turn — Browser Witness is non-negotiable.
+
+Before any apparent stop, dispatch `residual-scan`. If it returns work that fits the spirit of the original ask and is reachable from this session, expand the PRD and execute; only residuals genuinely out-of-spirit or out-of-reach are name-and-stop. A turn that ends with uncommitted changes, an open PRD slice, or unresolved mutables has not actually stopped — it has stalled the chain.
+
 The wasm artifact lives at `~/.claude/gm-tools/plugkit.wasm`; the spool watcher runs it. The watcher's own stdout/stderr is appended to `.gm/exec-spool/.watcher.log` — Read it to see plugkit's internal trace, dispatch timings, sweep actions, errors.
 
 ## Boot the spool watcher (first turn only)
@@ -84,11 +92,11 @@ The log is rotated at 10MB (older content moves to `.watcher.log.1`).
 
 ## The loop
 
-Dispatch `instruction` with empty body to get current-phase guidance + full state snapshot. Follow the `instruction` prose imperatively. Add PRD items via `prd-add` (JSON body), resolve via `prd-resolve` (id as body). Add mutables via `mutable-add`, resolve via `mutable-resolve` once `witness_evidence` is filled. Every resolve auto-fires `memorize-fire` so the evidence becomes recall-able.
+Add PRD items via `prd-add` (JSON body), resolve via `prd-resolve` (id as body). Add mutables via `mutable-add`, resolve via `mutable-resolve` once `witness_evidence` is filled — narrative resolution is rejected; only file:line, codesearch hit, or exec output snippet counts. Every `mutable-resolve` auto-fires memorize so the witness becomes recall-able next session.
 
 Resolve every entry in `mutables_pending` before transitioning. When the phase's exit condition is met, dispatch `transition` with the next phase name (or empty for auto-advance). Each transition response embeds `recall_hits` automatically — relevant prior memos surface without you asking.
 
-Stop when `next_phase_hint` is null or phase is `COMPLETE`.
+Stop only when `phase` is `COMPLETE` AND `residual-scan` returns empty AND the worktree is clean AND CI is green. Any of those false means the chain has not finished.
 
 ## Orchestrator verbs
 
@@ -104,8 +112,8 @@ Stop when `next_phase_hint` is null or phase is `COMPLETE`.
 
 ### Browser
 
-Dispatch `.gm/exec-spool/in/browser/<N>.txt` with raw JavaScript as the body. The wrapper spawns Chrome (managed profile at `<cwd>/.gm/browser-profile/`) and runs the JS via playwriter. Globals available inside the body: `page` (playwright Page), `snapshot` (accessibility snapshot), `screenshotWithAccessibilityLabels` (screenshot helper), `state` (per-session state object).
+The `browser` verb is the only sanctioned way to drive a live page. Do not reach for any other browser tool, library, or skill — the host owns the managed session and a parallel surface fragments witness state. Dispatch `.gm/exec-spool/in/browser/<N>.txt` with raw JavaScript as the body. The host runs Chrome under a project-scoped profile at `<cwd>/.gm/browser-profile/` (cookies/login persist per project) and exposes the body to four globals: `page` (the live page handle — `await page.goto(...)`, `await page.evaluate(...)`, etc.), `snapshot` (accessibility-tree snapshot), `screenshotWithAccessibilityLabels` (annotated screenshot helper), and `state` (a per-session object that persists across dispatches within the same session).
 
-Special commands (body starts with `session `): `session new`, `session list`, `session close <id>` pass through to playwriter directly.
+Special commands (body starts with `session `): `session new`, `session list`, `session close <id>` manage session lifecycle.
 
-Chrome is detected from system install paths; profile dir is project-scoped so cookies/login persist per project. The wrapper writes the managed gitignore block (which includes `.gm/browser-profile/`) on first launch.
+Required for any edit to code that runs in a browser — Browser Witness is non-negotiable. A `node test.js passes` does not substitute for a live `page.evaluate` asserting the invariant the edit was supposed to change.