@sparkleideas/claude-flow-patch 3.1.0-alpha.44.patch.5 → 3.1.0-alpha.44.patch.7

package/AGENTS.md

@@ -1,6 +1,6 @@
 # @sparkleideas/claude-flow-patch
 
-> Runtime patches for `@claude-flow/cli` v3.1.0-alpha.40, `ruvector`, and `ruv-swarm` v1.0.20
+> Runtime patches for `@claude-flow/cli` v3.1.0-alpha.41, `ruvector`, and `ruv-swarm` v1.0.20
 
 **Claude Code instructions are in [CLAUDE.md](CLAUDE.md). Project-specific instructions (defect index, workflows, policies) are in [README.md](README.md).**
 

package/CLAUDE.md

@@ -24,6 +24,7 @@ Runtime patches for `@claude-flow/cli` **v3.1.0-alpha.41**, `ruvector`, and `ruv
 - NEVER reuse a defect ID that was previously assigned to a different GitHub issue
 - ONE defect directory and ONE fix.py per GitHub issue -- do not combine multiple GitHub issues into one defect or split one GitHub issue across multiple defects
 - ALWAYS verify with `bash check-patches.sh` after applying
+- ALWAYS run `npm run preflight` before staging — the pre-commit hook (`hooks/pre-commit`) enforces this via `--check` mode
 - ALWAYS update ALL listing files when adding/removing a defect (see checklist)
 - Execution order is determined by the numeric prefix on each defect directory name. Dependencies between defects are expressed by assigning lower numbers to prerequisites.
 
@@ -38,7 +39,7 @@ lib/
   discover.mjs          # Dynamic discovery: scans patch/*/ → structured JSON
   categories.json       # Prefix-to-label mapping (one line per category)
 scripts/
-  update-docs.mjs       # Regenerates doc tables from discovery (npm run update-docs)
+  preflight.mjs         # Pre-commit sync: doc tables, versions, config (npm run preflight)
   upstream-log.mjs      # Show recent upstream releases (npm run upstream-log [count])
 patch/
   {NNN}-{PREFIX}-{NNN}-{slug}/    # NNN = 3-digit execution order
@@ -90,7 +91,13 @@ gh issue comment <NUMBER> --repo ruvnet/claude-flow --body "$(cat <<'EOF'
 
 Defect **{PREFIX}-{NNN}** in [claude-flow-patch](https://github.com/sparkling/claude-flow-patch).
 
+**Root cause:** <1-2 sentences explaining why the bug occurs at the code level>
+
 <What the patch does. Be specific. Include a table if multiple ops.>
+
+**Affected versions:** `@claude-flow/cli` 3.1.0-alpha.44 through current
+
+**Related issues:** #NNN, #NNN
 EOF
 )"
 ```
@@ -114,6 +121,12 @@ gh issue create --repo ruvnet/claude-flow \
 
 ## Files Affected
 - <dist/src/path/to/file.js>
+
+## Affected Versions
+`@claude-flow/cli` 3.1.0-alpha.44 through current
+
+## Related Issues
+- #NNN — <short description of relationship>
 EOF
 )"
 ```
@@ -123,6 +136,7 @@ Save the returned GitHub issue number for the defect README.md.
 ### Comment hygiene:
 
 - One comment per defect, describing the patch. No meta-commentary.
+- Every comment/body MUST include affected versions and related issues (use "None" if truly standalone).
 - If you need to replace a comment, delete the old one first (`gh api -X DELETE`).
 - Do not reference defect history, deletion/restoration, or internal decisions.
 
@@ -135,17 +149,17 @@ Save the returned GitHub issue number for the defect README.md.
 | DM | Daemon & Workers | 6 |
 | EM | Embeddings & HNSW | 2 |
 | GV | Ghost Vectors | 1 |
-| HK | Hooks | 3 |
+| HK | Hooks | 5 |
 | HW | Headless Worker | 4 |
 | IN | Intelligence | 1 |
 | MM | Memory Management | 1 |
 | NS | Memory Namespace | 3 |
 | RS | ruv-swarm | 1 |
 | RV | RuVector Intelligence | 3 |
-| SG | Settings Generator | 2 |
+| SG | Settings Generator | 4 |
 | UI | Display & Cosmetic | 2 |
 
-## All 31 Defects
+## All 35 Defects
 
 | ID | GitHub Issue | Severity |
 |----|-------------|----------|
@@ -162,9 +176,11 @@ Save the returned GitHub issue number for the defect README.md.
 | HK-001 | [#1155 post-edit hook records file_path as "unknown"](https://github.com/ruvnet/claude-flow/issues/1155) | Medium |
 | HK-002 | [#1058 MCP hook handlers are stubs that don't persist data](https://github.com/ruvnet/claude-flow/issues/1058) | High |
 | HK-003 | [#1158 hooks_metrics MCP handler returns hardcoded fake data](https://github.com/ruvnet/claude-flow/issues/1158) | High |
+| HK-004 | [#1175 hooks_session-start ignores daemon.autoStart from settings.json](https://github.com/ruvnet/claude-flow/issues/1175) | High |
+| HK-005 | [#1171 Multiple MCP servers start independent in-process daemons](https://github.com/ruvnet/claude-flow/issues/1171) | Critical |
 | HW-001 | [#1111 Headless workers hang — stdin pipe never closed](https://github.com/ruvnet/claude-flow/issues/1111) | Critical |
 | HW-002 | [#1112 Headless failures silently swallowed as success](https://github.com/ruvnet/claude-flow/issues/1112) | High |
-| HW-003 | [#1113 Worker scheduling intervals too aggressive](https://github.com/ruvnet/claude-flow/issues/1113) | High |
+| HW-003 | [#1113 Worker scheduling intervals too aggressive + settings ignored](https://github.com/ruvnet/claude-flow/issues/1113) | High |
 | IN-001 | [#1154 intelligence.cjs is a stub that doesn't actually learn](https://github.com/ruvnet/claude-flow/issues/1154) | Critical |
 | MM-001 | [#1152 Remove dead persistPath config option](https://github.com/ruvnet/claude-flow/issues/1152) | Low |
 | NS-001 | [#1123 Discovery ops default to wrong namespace](https://github.com/ruvnet/claude-flow/issues/1123) | Critical |
@@ -178,8 +194,10 @@ Save the returned GitHub issue number for the defect README.md.
 | SG-003 | [#1169 Init missing helpers for --dual, --minimal, hooks, and upgrade paths](https://github.com/ruvnet/claude-flow/issues/1169) | Critical |
 | UI-001 | [#1145 intelligence stats crashes on .toFixed()](https://github.com/ruvnet/claude-flow/issues/1145) | Critical |
 | UI-002 | [#1146 neural status shows "Not loaded"](https://github.com/ruvnet/claude-flow/issues/1146) | Low |
-| DM-006 | [#1114 No log rotation — headless execution logs grow unbounded](https://github.com/ruvnet/claude-flow/issues/1114) | Medium |
+| DM-006 | [#1114 No log rotation — logs grow unbounded](https://github.com/ruvnet/claude-flow/issues/1114) | Medium |
 | HW-004 | [#1117 runWithTimeout rejects but does not kill child process](https://github.com/ruvnet/claude-flow/issues/1117) | Medium |
+| SG-004 | [#1181 init wizard lacks parity with init](https://github.com/ruvnet/claude-flow/issues/1181) | High |
+| SG-005 | [#1177 add 'start all' subcommand to start everything at once](https://github.com/ruvnet/claude-flow/issues/1177) | Enhancement |
 <!-- GENERATED:defect-tables:end -->
 
 ---
@@ -305,6 +323,8 @@ Both are idempotent: skip if `new` already present, warn if `old` not found.
 | `SETTINGS_GEN` | `init/settings-generator.js` | @claude-flow/cli |
 | `HELPERS_GEN` | `init/helpers-generator.js` | @claude-flow/cli |
 | `EXECUTOR` | `init/executor.js` | @claude-flow/cli |
+| `INIT_CMD` | `commands/init.js` | @claude-flow/cli |
+| `START_CMD` | `commands/start.js` | @claude-flow/cli |
 | `ruvector_cli` | `bin/cli.js` | ruvector |
 
 To target a new file, add a variable to `lib/common.py` following the existing pattern.
@@ -313,7 +333,7 @@ To target a new file, add a variable to `lib/common.py` following the existing p
 
 ```bash
 # Regenerate all documentation from dynamic discovery
-npm run update-docs
+npm run preflight
 
 # Apply -- should show "Applied: ..."
 bash patch-all.sh --global
@@ -339,7 +359,7 @@ npm test
 - [ ] `patch/{PREFIX}-{NNN}-{slug}/sentinel` created with verification directives
 - [ ] Path variable added to `lib/common.py` (if targeting a new file)
 - [ ] If new category prefix: add one line to `lib/categories.json`
-- [ ] `npm run update-docs` regenerates all doc tables
+- [ ] `npm run preflight` regenerates all doc tables
 - [ ] `bash patch-all.sh` applies successfully
 - [ ] `bash patch-all.sh` is idempotent (0 applied on re-run)
 - [ ] `bash check-patches.sh` shows OK
@@ -355,7 +375,7 @@ Before removing any defect:
 1. Confirm the bug is genuinely fixed upstream or the patch is truly unreachable.
 2. Do NOT remove a defect just because a local workaround exists -- the MCP-level patch may still be needed.
 3. If removing, retire the defect ID permanently. Never reassign a deleted ID to a different GitHub issue.
-4. Run `npm run update-docs` to regenerate all documentation.
+4. Run `npm run preflight` to regenerate all documentation.
 
 ---
 
@@ -450,6 +470,31 @@ Two dependency chains exist:
 
 All other patches are independent.
 
+## Preflight & Pre-Commit Hook
+
+A git pre-commit hook at `hooks/pre-commit` runs automatically on every commit. It calls `npm run preflight:check` (read-only) and `npm test`. If anything is stale or tests fail, the commit is blocked.
+
+**Setup** (one-time, already done for this clone):
+```bash
+git config core.hooksPath hooks
+```
+
+**Before staging**, run:
+```bash
+npm run preflight    # Syncs doc tables, defect counts, version strings, config
+npm test             # Runs all tests
+```
+
+Then `git add -u` to stage the regenerated files.
+
+**What `preflight` syncs**:
+- Defect tables in README.md, CLAUDE.md, npm/README.md (from `patch/*/README.md`)
+- Defect counts in `npm/config.json` (from discovery)
+- `npm/config.json` version.current (from `package.json`)
+- Upstream baseline version in prose (from `npm/config.json` targets)
+
+Manual edits to generated sections (`<!-- GENERATED:*:begin/end -->`) will be overwritten.
+
 ## Key Design Decisions
 
 - **Idempotent**: `patch()` checks if `new` string is already present before replacing.

package/README.md

@@ -158,7 +158,7 @@ claude-flow-patch/
     discover.mjs         # Dynamic patch discovery — single source of truth
     categories.json      # Prefix-to-label mapping (e.g. HW → Headless Worker)
   scripts/
-    update-docs.mjs      # Regenerate doc tables from discover() output
+    preflight.mjs        # Pre-commit sync: doc tables, versions, config
   patch/
     {NNN}-{PREFIX}-{NNN}-{slug}/    # NNN = 3-digit execution order
       README.md          # Defect report: title, severity, root cause, fix
@@ -173,7 +173,7 @@ claude-flow-patch/
 ## Defect Index
 
 <!-- GENERATED:defect-index:begin -->
-31 defects across 13 categories.
+35 defects across 13 categories.
 
 ### CF -- Config & Doctor
 
@@ -191,7 +191,7 @@ claude-flow-patch/
 | [DM&#8209;003](patch/050-DM-003-macos-freemem/) | macOS freemem() always ~0% — workers blocked | Critical | [#1077](https://github.com/ruvnet/claude-flow/issues/1077) |
 | [DM&#8209;004](patch/060-DM-004-preload-worker-stub/) | Preload worker stub + missing from defaults | Enhancement | [#1139](https://github.com/ruvnet/claude-flow/issues/1139) |
 | [DM&#8209;005](patch/070-DM-005-consolidation-worker-stub/) | Consolidation worker stub (no decay/rebuild) | Enhancement | [#1140](https://github.com/ruvnet/claude-flow/issues/1140) |
-| [DM&#8209;006](patch/300-DM-006-log-rotation/) | No log rotation — headless execution logs grow unbounded | Medium | [#1114](https://github.com/ruvnet/claude-flow/issues/1114) |
+| [DM&#8209;006](patch/300-DM-006-log-rotation/) | No log rotation — logs grow unbounded | Medium | [#1114](https://github.com/ruvnet/claude-flow/issues/1114) |
 
 ### EM -- Embeddings & HNSW
 
@@ -213,6 +213,8 @@ claude-flow-patch/
 | [HK&#8209;001](patch/110-HK-001-post-edit-file-path/) | post-edit hook records file_path as "unknown" | Medium | [#1155](https://github.com/ruvnet/claude-flow/issues/1155) |
 | [HK&#8209;002](patch/120-HK-002-hooks-tools-stub/) | MCP hook handlers are stubs that don't persist data | High | [#1058](https://github.com/ruvnet/claude-flow/issues/1058) |
 | [HK&#8209;003](patch/130-HK-003-metrics-hardcoded/) | hooks_metrics MCP handler returns hardcoded fake data | High | [#1158](https://github.com/ruvnet/claude-flow/issues/1158) |
+| [HK&#8209;004](patch/135-HK-004-respect-daemon-autostart/) | hooks_session-start ignores daemon.autoStart from settings.json | High | [#1175](https://github.com/ruvnet/claude-flow/issues/1175) |
+| [HK&#8209;005](patch/137-HK-005-daemon-pid-guard/) | Multiple MCP servers start independent in-process daemons | Critical | [#1171](https://github.com/ruvnet/claude-flow/issues/1171) |
 
 ### HW -- Headless Worker
 
@@ -220,7 +222,7 @@ claude-flow-patch/
 |----|-------------|----------|--------------|
 | [HW&#8209;001](patch/140-HW-001-stdin-hang/) | Headless workers hang — stdin pipe never closed | Critical | [#1111](https://github.com/ruvnet/claude-flow/issues/1111) |
 | [HW&#8209;002](patch/150-HW-002-failures-swallowed/) | Headless failures silently swallowed as success | High | [#1112](https://github.com/ruvnet/claude-flow/issues/1112) |
-| [HW&#8209;003](patch/160-HW-003-aggressive-intervals/) | Worker scheduling intervals too aggressive | High | [#1113](https://github.com/ruvnet/claude-flow/issues/1113) |
+| [HW&#8209;003](patch/160-HW-003-aggressive-intervals/) | Worker scheduling intervals too aggressive + settings ignored | High | [#1113](https://github.com/ruvnet/claude-flow/issues/1113) |
 | [HW&#8209;004](patch/310-HW-004-runwithtimeout-orphan/) | runWithTimeout rejects but does not kill child process | Medium | [#1117](https://github.com/ruvnet/claude-flow/issues/1117) |
 
 ### IN -- Intelligence
@@ -263,6 +265,8 @@ claude-flow-patch/
 |----|-------------|----------|--------------|
 | [SG&#8209;001](patch/260-SG-001-init-settings/) | Init generates invalid settings | High | [#1150](https://github.com/ruvnet/claude-flow/issues/1150) |
 | [SG&#8209;003](patch/270-SG-003-init-helpers-all-paths/) | Init missing helpers for --dual, --minimal, hooks, and upgrade paths | Critical | [#1169](https://github.com/ruvnet/claude-flow/issues/1169) |
+| [SG&#8209;004](patch/320-SG-004-wizard-parity/) | init wizard lacks parity with init | High | [#1181](https://github.com/ruvnet/claude-flow/issues/1181) |
+| [SG&#8209;005](patch/330-SG-005-start-all-subcommand/) | add 'start all' subcommand to start everything at once | Enhancement | [#1177](https://github.com/ruvnet/claude-flow/issues/1177) |
 
 ### UI -- Display & Cosmetic
 

package/check-patches.sh

@@ -165,10 +165,14 @@ if [ -x "$SCRIPT_DIR/patch-all.sh" ]; then
   if [[ -n "$TARGET_DIR" ]]; then REAPPLY_ARGS+=(--target "$TARGET_DIR"); fi
   bash "$SCRIPT_DIR/patch-all.sh" "${REAPPLY_ARGS[@]}"
   echo ""
-  echo "[PATCHES] Auto-reapplied. Restarting daemon..."
+  echo "[PATCHES] Auto-reapplied. Stopping existing daemons..."
   npx @claude-flow/cli@latest daemon stop 2>/dev/null
+  # Fallback: kill by PID file if daemon stop missed an orphan (project-scoped, not global)
+  _pid=$(cat .claude-flow/daemon.pid 2>/dev/null)
+  if [ -n "$_pid" ]; then kill "$_pid" 2>/dev/null || true; rm -f .claude-flow/daemon.pid; fi
+  sleep 1
   npx @claude-flow/cli@latest daemon start 2>/dev/null
-  echo "[PATCHES] Daemon restarted with patched code."
+  echo "[PATCHES] Daemon restarted in background (PID: $(cat .claude-flow/daemon.pid 2>/dev/null || echo 'unknown'))"
   echo ""
 else
   echo "[PATCHES] ERROR: patch-all.sh not found at $SCRIPT_DIR"

package/lib/common.py

@@ -83,6 +83,7 @@ HELPERS_GEN = init + "/helpers-generator.js" if init else ""
 EXECUTOR = init + "/executor.js" if init else ""
 TYPES = init + "/types.js" if init else ""
 INIT_CMD = commands + "/init.js" if commands else ""
+START_CMD = commands + "/start.js" if commands else ""
 
 # Source helpers (shipped with package, copied by writeHelpers when source dir found)
 _pkg_root = os.path.dirname(os.path.dirname(base)) if base else ""

package/package.json

@@ -1,15 +1,16 @@
 {
   "name": "@sparkleideas/claude-flow-patch",
-  "version": "3.1.0-alpha.44.patch.5",
+  "version": "3.1.0-alpha.44.patch.7",
   "description": "Patch toolkit for @claude-flow/cli init/runtime defects with verify and post-init repair commands",
   "scripts": {
+    "preflight": "node scripts/preflight.mjs",
+    "preflight:check": "node scripts/preflight.mjs --check",
+    "prepublishOnly": "npm run preflight && npm test",
     "test": "node --test tests/*.test.mjs",
     "package": "bash npm/package.sh",
     "package:dry": "bash npm/package.sh --dry-run",
     "publish:npm": "bash npm/publish.sh",
     "publish:dry": "bash npm/publish.sh --dry-run",
-    "update-docs": "node scripts/update-docs.mjs",
-    "update-docs:check": "node scripts/update-docs.mjs --check",
     "upstream-log": "node scripts/upstream-log.mjs",
     "guidance:analyze": "guidance analyze",
     "guidance:optimize": "guidance autopilot --once --apply --source manual",

package/patch/135-HK-004-respect-daemon-autostart/README.md

@@ -0,0 +1,11 @@
+# HK-004: hooks_session-start ignores daemon.autoStart from settings.json
+**Severity**: High
+**GitHub**: [#1175](https://github.com/ruvnet/claude-flow/issues/1175)
+## Root Cause
+`hooks_session-start` handler in `hooks-tools.js` (line ~1216) determines daemon auto-start solely from the MCP call parameter: `const shouldStartDaemon = params.startDaemon !== false;`. It never reads `claudeFlow.daemon.autoStart` from `.claude/settings.json`. Setting `autoStart: false` in settings.json has no effect — the daemon always starts on session-start.
+## Fix
+Wrap the `shouldStartDaemon` assignment in an IIFE that first checks the MCP parameter, then reads settings.json. If `claudeFlow.daemon.autoStart === false`, returns false. Falls back to true on any read/parse error.
+## Files Patched
+- mcp-tools/hooks-tools.js
+## Ops
+1 op in fix.py

package/patch/135-HK-004-respect-daemon-autostart/fix.py

@@ -0,0 +1,14 @@
+# HK-004: hooks_session-start ignores daemon.autoStart from settings.json
+# GitHub: #1175
+patch("HK-004: respect daemon autoStart setting",
+    MCP_HOOKS,
+    "        const shouldStartDaemon = params.startDaemon !== false;",
+    """        const shouldStartDaemon = (() => {
+            if (params.startDaemon === false) return false;
+            try {
+                const sp = join(process.cwd(), '.claude', 'settings.json');
+                const s = JSON.parse(readFileSync(sp, 'utf-8'));
+                if (s?.claudeFlow?.daemon?.autoStart === false) return false;
+            } catch { /* no settings or unreadable — default to true */ }
+            return true;
+        })();""")

package/patch/135-HK-004-respect-daemon-autostart/sentinel

@@ -0,0 +1 @@
+grep "claudeFlow?.daemon?.autoStart" mcp-tools/hooks-tools.js

package/patch/137-HK-005-daemon-pid-guard/README.md

@@ -0,0 +1,11 @@
+# HK-005: Multiple MCP servers start independent in-process daemons
+**Severity**: Critical
+**GitHub**: [#1171](https://github.com/ruvnet/claude-flow/issues/1171)
+## Root Cause
+`hooks_session-start` calls `startDaemon()` (worker-daemon.js) which creates an in-process `WorkerDaemon` singleton per Node.js process. Each MCP server is a separate process with its own singleton — no cross-process coordination. The CLI background daemon path (`daemon.js`) has PID-file coordination but the MCP hook path bypasses it entirely. Result: N MCP servers per project = N daemon instances = N × 6 workers.
+## Fix
+Add PID-file guard to the MCP hook path using `.claude-flow/daemon.pid` (same file the CLI path uses). Before calling `startDaemon()`, check PID file: if a different process owns it and is alive, skip (reuse). If the PID is our own process or stale, proceed and overwrite. No cleanup on session-end — stale PIDs self-heal via `kill(pid, 0)` on next start. The PID file becomes a universal one-daemon-per-project lock across both MCP and CLI paths.
+## Files Patched
+- mcp-tools/hooks-tools.js
+## Ops
+2 ops in fix.py

package/patch/137-HK-005-daemon-pid-guard/fix.py

@@ -0,0 +1,53 @@
+# HK-005: Multiple MCP servers start independent in-process daemons
+# No cross-process coordination on the hooks_session-start path.
+# GitHub: #1171
+
+# Op 1: Add PID-file guard before startDaemon()
+patch("HK-005a: cross-process daemon PID guard",
+    MCP_HOOKS,
+    """        // Auto-start daemon if enabled
+        let daemonStatus = { started: false };
+        if (shouldStartDaemon) {
+            try {
+                // Dynamic import to avoid circular dependencies
+                const { startDaemon } = await import('../services/worker-daemon.js');
+                const daemon = await startDaemon(process.cwd());""",
+    """        // Auto-start daemon if enabled
+        let daemonStatus = { started: false };
+        if (shouldStartDaemon) {
+            try {
+                // HK-005: PID-file guard — one daemon per project across processes
+                const _pidDir = join(process.cwd(), '.claude-flow');
+                const _pidPath = join(_pidDir, 'daemon.pid');
+                let _skipDaemon = false;
+                try {
+                    const _xPid = parseInt(readFileSync(_pidPath, 'utf-8').trim(), 10);
+                    if (!isNaN(_xPid) && _xPid !== process.pid) {
+                        try { process.kill(_xPid, 0); _skipDaemon = true; daemonStatus = { started: true, pid: _xPid, reused: true }; }
+                        catch { /* stale PID from dead process — proceed */ }
+                    }
+                } catch { /* no PID file — proceed */ }
+                if (!_skipDaemon) {
+                // Dynamic import to avoid circular dependencies
+                const { startDaemon } = await import('../services/worker-daemon.js');
+                const daemon = await startDaemon(process.cwd());""")
+
+# Op 2: Write PID after successful start + close guard block
+patch("HK-005b: write PID after daemon start",
+    MCP_HOOKS,
+    """                const status = daemon.getStatus();
+                daemonStatus = {
+                    started: true,
+                    pid: status.pid,
+                };""",
+    """                const status = daemon.getStatus();
+                // HK-005: Write PID so other processes detect this daemon
+                try {
+                    if (!existsSync(_pidDir)) { mkdirSync(_pidDir, { recursive: true }); }
+                    writeFileSync(_pidPath, String(status.pid || process.pid));
+                } catch { /* best-effort */ }
+                daemonStatus = {
+                    started: true,
+                    pid: status.pid,
+                };
+                } // end HK-005 guard""")

package/patch/137-HK-005-daemon-pid-guard/sentinel

@@ -0,0 +1,2 @@
+grep "HK-005: PID-file guard" mcp-tools/hooks-tools.js
+grep "HK-005: Write PID" mcp-tools/hooks-tools.js

package/patch/160-HW-003-aggressive-intervals/README.md

@@ -1,11 +1,11 @@
-# HW-003: Worker scheduling intervals too aggressive
+# HW-003: Worker scheduling intervals too aggressive + settings ignored
 **Severity**: High
 **GitHub**: [#1113](https://github.com/ruvnet/claude-flow/issues/1113)
 ## Root Cause
-`DEFAULT_WORKERS` uses pre-headless intervals (audit: 10m, optimize: 15m, testgaps: 20m). ADR-020 specifies longer intervals (30/60/60m) for headless workers that invoke Claude.
+`DEFAULT_WORKERS` uses pre-headless intervals (audit: 10m, optimize: 15m, testgaps: 20m). ADR-020 specifies longer intervals (30/60/60m) for headless workers that invoke Claude. Additionally, `daemon.schedules` from `.claude/settings.json` is never read — user-configured intervals are completely ignored.
 ## Fix
-Align intervals to ADR-020: audit 30m, optimize 60m, testgaps 60m.
+(A) Align hardcoded intervals to ADR-020: audit 30m, optimize 60m, testgaps 60m. (B) After setting default workers, read `claudeFlow.daemon.schedules` from `.claude/settings.json` and merge user-configured intervals/enabled flags into matching workers. Supports string formats ("1h", "30m", "10s") and raw milliseconds.
 ## Files Patched
 - services/worker-daemon.js
 ## Ops
-3 ops in fix.py
+4 ops in fix.py

package/patch/160-HW-003-aggressive-intervals/fix.py

@@ -14,3 +14,39 @@ patch("3: testgaps 60m",
     WD,
     "type: 'testgaps', intervalMs: 20 * 60 * 1000",
     "type: 'testgaps', intervalMs: 60 * 60 * 1000")
+
+# HW-003 extension: Read daemon.schedules from .claude/settings.json
+# and merge user-configured intervals into worker defaults
+patch("3: settings-driven intervals",
+    WD,
+    "            workers: config?.workers ?? DEFAULT_WORKERS,",
+    """            workers: (() => {
+                const base = config?.workers ?? DEFAULT_WORKERS;
+                try {
+                    const sp = join(projectRoot, '.claude', 'settings.json');
+                    const s = JSON.parse(readFileSync(sp, 'utf-8'));
+                    const schedules = s?.claudeFlow?.daemon?.schedules;
+                    if (!schedules || typeof schedules !== 'object') return base;
+                    const parseInterval = (v) => {
+                        if (typeof v === 'number') return v;
+                        if (typeof v !== 'string') return null;
+                        const m = v.match(/^(\\d+(?:\\.\\d+)?)\\s*(ms|s|m|h)$/i);
+                        if (!m) return null;
+                        const n = parseFloat(m[1]);
+                        switch (m[2].toLowerCase()) {
+                            case 'ms': return n;
+                            case 's': return n * 1000;
+                            case 'm': return n * 60 * 1000;
+                            case 'h': return n * 3600 * 1000;
+                            default: return null;
+                        }
+                    };
+                    return base.map(w => {
+                        const sched = schedules[w.type];
+                        if (!sched) return w;
+                        const iv = parseInterval(sched.interval ?? sched.intervalMs);
+                        const en = typeof sched.enabled === 'boolean' ? sched.enabled : w.enabled;
+                        return { ...w, ...(iv !== null ? { intervalMs: iv } : {}), enabled: en };
+                    });
+                } catch { return base; }
+            })(),""")

package/patch/160-HW-003-aggressive-intervals/sentinel

@@ -1 +1,3 @@
-grep "30 * 60 * 1000" services/worker-daemon.js
+grep "intervalMs: 30" services/worker-daemon.js
+grep "parseInterval" services/worker-daemon.js
+grep "daemon?.schedules" services/worker-daemon.js

package/patch/300-DM-006-log-rotation/README.md

@@ -1,11 +1,12 @@
-# DM-006: No log rotation — headless execution logs grow unbounded
+# DM-006: No log rotation — logs grow unbounded
 **Severity**: Medium
 **GitHub**: [#1114](https://github.com/ruvnet/claude-flow/issues/1114)
 ## Root Cause
-`logExecution()` in `headless-worker-executor.js` creates 2-3 log files per worker run (~75 KB each) but has zero cleanup. No rotation, no max file count, no TTL. At current daemon intervals this accumulates ~23 MB/day, ~702 MB/month.
+(Headless) `logExecution()` in `headless-worker-executor.js` creates 2-3 log files per worker run (~75 KB each) but has zero cleanup. No rotation, no max file count, no TTL. At current daemon intervals this accumulates ~23 MB/day, ~702 MB/month. (Main) `startBackgroundDaemon()` in `daemon.js` opens `daemon.log` in append mode and never truncates. A single long-running daemon can grow daemon.log to 100+ GB.
 ## Fix
-(A) Add `unlinkSync` and `statSync` to the ESM import. (B) Call `cleanupOldLogs()` from `ensureLogDir()` so cleanup runs on each execution cycle. (C) Add `cleanupOldLogs()` method: removes `.log` files older than 7 days or beyond a 500-file cap, keeping newest files.
+(A) Add `unlinkSync` and `statSync` to the headless executor ESM import. (B) Call `cleanupOldLogs()` from `ensureLogDir()` so cleanup runs on each headless execution cycle. (C) Add `cleanupOldLogs()` method: removes `.log` files older than 7 days or beyond a 500-file cap, keeping newest files. (D) In `daemon.js`, before opening `daemon.log` for append, check its size with `fs.statSync()` and rotate to `daemon.log.1` if > 50MB.
 ## Files Patched
 - services/headless-worker-executor.js
+- commands/daemon.js
 ## Ops
-3 ops in fix.py
+4 ops in fix.py

package/patch/300-DM-006-log-rotation/fix.py

@@ -56,3 +56,17 @@ patch("DM-006c: add cleanupOldLogs method",
         catch { /* ignore cleanup errors */ }
     }
     logExecution(executionId, type, content) {""")
+
+# DM-006 extension: Main daemon.log rotation in daemon.js
+# Before opening daemon.log for append, check size and rotate if > 50MB
+patch("DM-006d: main daemon.log rotation",
+    DJ,
+    "    const logFile = join(logsDir, 'daemon.log');",
+    """    const logFile = join(logsDir, 'daemon.log');
+    // Rotate main daemon.log if > 50MB
+    try {
+        const logStat = fs.statSync(logFile);
+        if (logStat.size > 50 * 1024 * 1024) {
+            fs.renameSync(logFile, logFile + '.1');
+        }
+    } catch { /* file doesn't exist yet or stat failed — ignore */ }""")

package/patch/300-DM-006-log-rotation/sentinel

@@ -1 +1,2 @@
 grep "cleanupOldLogs" services/headless-worker-executor.js
+grep "Rotate main daemon.log" commands/daemon.js

package/patch/320-SG-004-wizard-parity/README.md

@@ -0,0 +1,32 @@
+# SG-004: init wizard lacks parity with init
+
+**Severity**: High
+**GitHub**: [#1181](https://github.com/ruvnet/claude-flow/issues/1181)
+
+## Root Cause
+
+The `wizardCommand.action` in `commands/init.js` was implemented as a
+standalone code path that diverges from the parent `initAction`. It skips the
+already-initialized guard, ignores `--force`, `--start-all`, `--start-daemon`,
+`--codex`, and `--dual` flags, and never shows "Next steps" hints. The wizard
+is conceptually "init with interactive value selection" but behaves as a
+completely separate command.
+
+## Fix
+
+Four ops bring the wizard to full parity with `init`:
+
+| Op | What it does |
+|----|-------------|
+| SG-004a | Adds the already-initialized guard before prompts + passes `--force` to options |
+| SG-004b | Adds `--codex`/`--dual` handling after executeInit succeeds |
+| SG-004c | Adds `--start-all`/`--start-daemon` service startup + "Next steps" hints |
+| SG-004d | Fixes catch block — catches errors cleanly instead of re-throwing |
+
+## Files Patched
+
+- `commands/init.js`
+
+## Ops
+
+4 ops in fix.py

package/patch/320-SG-004-wizard-parity/fix.py

@@ -0,0 +1,133 @@
+# SG-004: init wizard lacks parity with init
+# GitHub: #1181
+#
+# The wizard was implemented as a standalone code path that ignores
+# --force, --start-all, --start-daemon, --codex, --dual, and skips
+# the already-initialized guard and "Next steps" hints.
+#
+# 4 ops: init-guard + force (a), codex/dual (b), start-all + next-steps (c),
+#         catch-block error handling (d)
+
+# Op 1: Add already-initialized guard + pass --force to executeInit options
+patch("SG-004a: wizard init-guard + --force",
+    INIT_CMD,
+    """        try {
+            // Start with base options
+            const options = { ...DEFAULT_INIT_OPTIONS, targetDir: ctx.cwd };""",
+    """        try {
+            // SG-004: Check if already initialized (respects --force)
+            const force = ctx.flags.force;
+            const initialized = isInitialized(ctx.cwd);
+            const hasExisting = initialized.claude || initialized.claudeFlow;
+            if (hasExisting && !force) {
+                output.printWarning('Claude Flow appears to be already initialized');
+                if (initialized.claude) output.printInfo('  Found: .claude/settings.json');
+                if (initialized.claudeFlow) output.printInfo('  Found: .claude-flow/config.yaml');
+                output.printInfo('Use --force to reinitialize');
+                const proceed = await confirm({
+                    message: 'Do you want to reinitialize? This will overwrite existing configuration.',
+                    default: false,
+                });
+                if (!proceed) {
+                    return { success: true, message: 'Wizard cancelled' };
+                }
+            }
+            // Start with base options
+            const options = { ...DEFAULT_INIT_OPTIONS, targetDir: ctx.cwd, force: ctx.flags.force };""")
+
+# Op 2: Add --codex / --dual handling after executeInit succeeds
+patch("SG-004b: wizard --codex/--dual support",
+    INIT_CMD,
+    """            spinner.succeed('Setup complete!');
+            // Initialize embeddings if enabled
+            let embeddingsInitialized = false;""",
+    """            spinner.succeed('Setup complete!');
+            // SG-004: Respect --codex / --dual in wizard
+            const codexMode = ctx.flags.codex;
+            const dualMode = ctx.flags.dual;
+            if (codexMode || dualMode) {
+                try {
+                    output.writeln(output.dim('  Initializing Codex integration...'));
+                    await initCodexAction(ctx, { codexMode, dualMode, force: ctx.flags.force, minimal: false, full: false });
+                } catch (err) {
+                    output.printWarning(`Codex initialization: ${err instanceof Error ? err.message : String(err)}`);
+                }
+            }
+            // Initialize embeddings if enabled
+            let embeddingsInitialized = false;""")
+
+# Op 3: Add --start-all / --start-daemon + "Next steps" hints before final return
+patch("SG-004c: wizard --start-all + next-steps",
+    INIT_CMD,
+    """            });
+            return { success: true, data: result };
+        }
+        catch (error) {
+            if (error instanceof Error && error.message === 'User cancelled') {""",
+    """            });
+            // SG-004: Respect --start-all / --start-daemon in wizard
+            const startAll = ctx.flags['start-all'] || ctx.flags.startAll;
+            const startDaemon = ctx.flags['start-daemon'] || ctx.flags.startDaemon || startAll;
+            if (startDaemon || startAll) {
+                output.writeln();
+                output.printInfo('Starting services...');
+                const { execSync } = await import('child_process');
+                if (startAll) {
+                    try {
+                        output.writeln(output.dim('  Initializing memory database...'));
+                        execSync('npx @claude-flow/cli@latest memory init 2>/dev/null', {
+                            stdio: 'pipe', cwd: ctx.cwd, timeout: 30000
+                        });
+                        output.writeln(output.success('  \\u2713 Memory initialized'));
+                    } catch { output.writeln(output.dim('  Memory database already exists')); }
+                }
+                if (startDaemon) {
+                    try {
+                        output.writeln(output.dim('  Starting daemon...'));
+                        execSync('npx @claude-flow/cli@latest daemon start 2>/dev/null &', {
+                            stdio: 'pipe', cwd: ctx.cwd, timeout: 10000
+                        });
+                        output.writeln(output.success('  \\u2713 Daemon started'));
+                    } catch { output.writeln(output.warning('  Daemon may already be running')); }
+                }
+                if (startAll) {
+                    try {
+                        output.writeln(output.dim('  Initializing swarm...'));
+                        execSync('npx @claude-flow/cli@latest swarm init --topology hierarchical 2>/dev/null', {
+                            stdio: 'pipe', cwd: ctx.cwd, timeout: 30000
+                        });
+                        output.writeln(output.success('  \\u2713 Swarm initialized'));
+                    } catch { output.writeln(output.dim('  Swarm initialization skipped')); }
+                }
+                output.writeln();
+                output.printSuccess('All services started');
+            }
+            else {
+                output.writeln(output.bold('Next steps:'));
+                output.printList([
+                    `Run ${output.highlight('claude-flow daemon start')} to start background workers`,
+                    `Run ${output.highlight('claude-flow memory init')} to initialize memory database`,
+                    `Run ${output.highlight('claude-flow swarm init')} to initialize a swarm`,
+                    `Or re-run with ${output.highlight('--start-all')} to do all of the above`,
+                ]);
+            }
+            return { success: true, data: result };
+        }
+        catch (error) {
+            if (error instanceof Error && error.message === 'User cancelled') {""")
+
+# Op 4: Fix wizard catch block — re-throws instead of clean error message
+# spinner is declared inside the try block so we can't call spinner.fail() here
+patch("SG-004d: wizard catch block handles errors cleanly",
+    INIT_CMD,
+    """            if (error instanceof Error && error.message === 'User cancelled') {
+                output.printInfo('Setup cancelled');
+                return { success: true };
+            }
+            throw error;""",
+    """            if (error instanceof Error && error.message === 'User cancelled') {
+                output.printInfo('Setup cancelled');
+                return { success: true };
+            }
+            output.printError(`Failed to initialize: ${error instanceof Error ? error.message : String(error)}`);
+            return { success: false, exitCode: 1 };""")

package/patch/320-SG-004-wizard-parity/sentinel

@@ -0,0 +1 @@
+grep "SG-004" commands/init.js

package/patch/330-SG-005-start-all-subcommand/README.md

@@ -0,0 +1,29 @@
+# SG-005: add 'start all' subcommand to start everything at once
+
+**Severity**: Enhancement
+**GitHub**: [#1177](https://github.com/ruvnet/claude-flow/issues/1177)
+
+## Root Cause
+
+There is no single command to start the full Claude Flow stack (memory +
+daemon + swarm + MCP) on an already-initialized project. `claude-flow start`
+only initializes the swarm and MCP server. Users must run `memory init`,
+`daemon start`, and `start` separately, or re-run `init --start-all` which
+also re-creates project files.
+
+## Fix
+
+Add an `allCommand` subcommand to the `start` command so that
+`claude-flow start all` initializes memory, starts the daemon, then runs the
+normal `startAction` (swarm + MCP + health checks).
+
+Also adds the subcommand to the `subcommands` array and a corresponding
+example entry.
+
+## Files Patched
+
+- `commands/start.js`
+
+## Ops
+
+2 ops in fix.py

package/patch/330-SG-005-start-all-subcommand/fix.py

@@ -0,0 +1,50 @@
+# SG-005: add 'start all' subcommand to start everything at once
+# GitHub: #1177
+
+patch("SG-005a: add allCommand subcommand definition",
+    START_CMD,
+    """// Quick start subcommand
+const quickCommand = {""",
+    """// Start-all subcommand — SG-005
+const allCommand = {
+    name: 'all',
+    aliases: ['everything'],
+    description: 'Start memory, daemon, swarm, and MCP server',
+    action: async (ctx) => {
+        // Check initialization
+        if (!isInitialized(ctx.cwd)) {
+            output.printError('Claude Flow is not initialized in this directory');
+            output.printInfo('Run "claude-flow init" first, or use "claude-flow start quick"');
+            return { success: false, exitCode: 1 };
+        }
+        output.writeln();
+        output.writeln(output.bold('Starting all Claude Flow services'));
+        output.writeln();
+        const { execSync } = await import('child_process');
+        // Step 1: Initialize memory
+        try {
+            output.writeln(output.dim('  Initializing memory database...'));
+            execSync('npx @claude-flow/cli@latest memory init 2>/dev/null', {
+                stdio: 'pipe', cwd: ctx.cwd, timeout: 30000
+            });
+            output.writeln('  \\u2713 Memory initialized');
+        } catch { output.writeln('  Memory database already exists'); }
+        // Step 2: Start daemon
+        try {
+            output.writeln(output.dim('  Starting daemon...'));
+            execSync('npx @claude-flow/cli@latest daemon start 2>/dev/null &', {
+                stdio: 'pipe', cwd: ctx.cwd, timeout: 10000
+            });
+            output.writeln('  \\u2713 Daemon started');
+        } catch { output.writeln('  Daemon may already be running'); }
+        // Step 3: Start swarm + MCP via normal startAction
+        return startAction(ctx);
+    }
+};
+// Quick start subcommand
+const quickCommand = {""")
+
+patch("SG-005b: register allCommand in subcommands array and add example",
+    START_CMD,
+    """    subcommands: [stopCommand, restartCommand, quickCommand],""",
+    """    subcommands: [stopCommand, restartCommand, quickCommand, allCommand],""")

package/patch/330-SG-005-start-all-subcommand/sentinel

@@ -0,0 +1 @@
+grep "allCommand" commands/start.js

package/scripts/{update-docs.mjs → preflight.mjs}

@@ -1,9 +1,10 @@
 #!/usr/bin/env node
-// scripts/update-docs.mjs — Regenerate documentation from dynamic patch discovery.
-// Updates: README.md, CLAUDE.md, npm/README.md, npm/config.json
+// scripts/preflight.mjs — Pre-commit/pre-publish consistency check.
+// Syncs: doc tables, defect counts, version strings across all files.
+// Source of truth: package.json (version), npm/config.json (targets), patch/*/ (defects).
 //
-// Usage: node scripts/update-docs.mjs [--check]
-//   --check  Exit 1 if docs are out of date (for CI), don't write.
+// Usage: node scripts/preflight.mjs [--check]
+//   --check  Exit 1 if anything is out of date (for hooks/CI), don't write.
 
 import { readFileSync, writeFileSync } from 'node:fs';
 import { resolve, dirname } from 'node:path';
@@ -17,6 +18,14 @@ const checkOnly = process.argv.includes('--check');
 const data = discover();
 const { patches, categories, stats } = data;
 
+// ── Sources of truth ──
+
+const pkgJson = JSON.parse(readFileSync(resolve(ROOT, 'package.json'), 'utf-8'));
+const configJson = JSON.parse(readFileSync(resolve(ROOT, 'npm', 'config.json'), 'utf-8'));
+const pkgVersion = pkgJson.version;
+const cliTarget = configJson.targets['@claude-flow/cli'];
+const swarmTarget = configJson.targets['ruv-swarm'];
+
 // ── Helpers ──
 
 /** Group patches by prefix, preserving sort order. */
@@ -57,6 +66,20 @@ function replaceMarkerSection(filePath, markerName, newContent) {
   return true;
 }
 
+/**
+ * Replace all occurrences of a version string in a file.
+ * Returns true if content changed.
+ */
+function syncVersionInFile(filePath, oldVersion, newVersion, label) {
+  if (oldVersion === newVersion) return false;
+  const text = readFileSync(filePath, 'utf-8');
+  if (!text.includes(oldVersion)) return false;
+  const updated = text.replaceAll(oldVersion, newVersion);
+  if (updated === text) return false;
+  if (!checkOnly) writeFileSync(filePath, updated);
+  return true;
+}
+
 // ── Generate README.md defect index ──
 
 function generateReadmeIndex() {
@@ -142,13 +165,21 @@ function updateNpmReadme() {
   );
 }
 
-// ── Update npm/config.json counts ──
+// ── Sync npm/config.json (version + defect counts) ──
 
 function updateNpmConfig() {
   const filePath = resolve(ROOT, 'npm', 'config.json');
   const config = JSON.parse(readFileSync(filePath, 'utf-8'));
 
   let changed = false;
+
+  // Sync version.current from package.json
+  if (config.version?.current !== pkgVersion) {
+    config.version.current = pkgVersion;
+    changed = true;
+  }
+
+  // Sync defect counts from discovery
   if (config.defects?.total !== stats.total) {
     config.defects.total = stats.total;
     changed = true;
@@ -168,41 +199,51 @@ function updateNpmConfig() {
 
 let anyChanged = false;
 
-const readmeChanged = replaceMarkerSection(
-  resolve(ROOT, 'README.md'),
-  'defect-index',
-  generateReadmeIndex()
-);
-if (readmeChanged) {
+function report(changed, label) {
+  if (!changed) return;
   anyChanged = true;
-  console.log(checkOnly ? 'STALE: README.md' : 'Updated: README.md');
+  console.log(checkOnly ? `STALE: ${label}` : `Updated: ${label}`);
 }
 
-const claudeChanged = replaceMarkerSection(
-  resolve(ROOT, 'CLAUDE.md'),
-  'defect-tables',
-  generateClaudeTables()
+// 1. Doc tables (marker-based sections)
+report(
+  replaceMarkerSection(resolve(ROOT, 'README.md'), 'defect-index', generateReadmeIndex()),
+  'README.md (defect index)'
 );
-if (claudeChanged) {
-  anyChanged = true;
-  console.log(checkOnly ? 'STALE: CLAUDE.md' : 'Updated: CLAUDE.md');
-}
-
-const npmReadmeChanged = updateNpmReadme();
-if (npmReadmeChanged) {
-  anyChanged = true;
-  console.log(checkOnly ? 'STALE: npm/README.md' : 'Updated: npm/README.md');
-}
-
-const npmConfigChanged = updateNpmConfig();
-if (npmConfigChanged) {
-  anyChanged = true;
-  console.log(checkOnly ? 'STALE: npm/config.json' : 'Updated: npm/config.json');
+report(
+  replaceMarkerSection(resolve(ROOT, 'CLAUDE.md'), 'defect-tables', generateClaudeTables()),
+  'CLAUDE.md (defect tables)'
+);
+report(updateNpmReadme(), 'npm/README.md (defect list)');
+
+// 2. Config sync (version + counts)
+report(updateNpmConfig(), 'npm/config.json (version/counts)');
+
+// 3. Upstream baseline version in prose (sync from npm/config.json targets)
+// Find any stale version strings and replace with current targets.
+// We scan for the pattern v?X.Y.Z-alpha.N and replace if it doesn't match config.
+const versionFiles = ['README.md', 'CLAUDE.md', 'npm/README.md', 'AGENTS.md'];
+for (const file of versionFiles) {
+  const filePath = resolve(ROOT, file);
+  let text;
+  try { text = readFileSync(filePath, 'utf-8'); } catch { continue; }
+
+  let updated = text;
+
+  // Sync @claude-flow/cli version references
+  // Match patterns like **v3.1.0-alpha.NN** or `3.1.0-alpha.NN` or @3.1.0-alpha.NN
+  const cliRe = /(?<=[@`*v])3\.1\.0-alpha\.\d+/g;
+  updated = updated.replace(cliRe, cliTarget);
+
+  if (updated !== text) {
+    if (!checkOnly) writeFileSync(filePath, updated);
+    report(true, `${file} (upstream baseline)`);
+  }
 }
 
 if (!anyChanged) {
-  console.log('All docs are up to date.');
+  console.log('All files are up to date.');
 } else if (checkOnly) {
-  console.log('\nDocs are out of date. Run: npm run update-docs');
+  console.log('\nFiles are out of date. Run: npm run preflight');
   process.exit(1);
 }