specweave 0.22.3 → 0.22.5

package/CLAUDE.md

@@ -1,1266 +1,497 @@
 # SpecWeave - Development Guide
 
 **Project**: SpecWeave - Spec-Driven Development Framework
-**Type**: Open Source NPM Package (TypeScript CLI)
+**Type**: TypeScript CLI (NPM Package)
 **Repository**: https://github.com/anton-abyzov/specweave
-**Website**: https://spec-weave.com
 
-This CLAUDE.md is for **contributors to SpecWeave itself**, not users of SpecWeave.
-Users receive a different CLAUDE.md via the template system.
+For **contributors to SpecWeave itself** (not users).
 
 ---
 
-## 🔍 CRITICAL FINDING: Claude Code Marketplace Directory vs Symlink Issue
+## 🚨 CRITICAL SAFETY RULES
 
-**⚠️ RECENTLY DISCOVERED (2025-11-18) - AFFECTS ALL SPECWEAVE CONTRIBUTORS ⚠️**
+### 1. Symlink Setup (MANDATORY for Contributors)
 
-### The Issue
-
-Claude Code's hook execution system expects plugins at `~/.claude/plugins/marketplaces/specweave/`, but this path can be **either a directory OR a symlink**. The difference is critical:
-
-| Type | Symbol | Behavior | Development Impact |
-|------|--------|----------|-------------------|
-| **Directory** | `drwxr-xr-x` | Stale copy from marketplace update | ❌ Hooks fail with "No such file or directory" |
-| **Symlink** | `lrwxr-xr-x` | Live reference to repository | ✅ Hooks work, changes immediately reflected |
-
-### Why This Matters
-
-**Hooks ONLY work reliably with a symlink setup.** If the marketplace path is a directory:
-- ❌ Post-task-completion hooks fail randomly
-- ❌ Status line doesn't update
-- ❌ Living docs don't sync
-- ❌ You waste hours debugging "No such file or directory" errors
-- ❌ Changes to hooks in your repo are NOT reflected until marketplace update
-
-### How the Problem Occurs
-
-1. **Initial setup**: You create a symlink (correct)
-2. **Marketplace update**: Running `claude plugin marketplace update specweave` MAY replace the symlink with a directory copy
-3. **Silent failure**: Hooks start failing but no clear error message points to symlink issue
-
-### The Fix
+**Problem**: Claude Code executes hooks from `~/.claude/plugins/marketplaces/specweave/`. If this is a **directory** (not symlink), hooks fail with "No such file or directory".
 
+**Fix**:
 ```bash
-# ALWAYS verify your setup is a symlink:
+# Verify symlink exists:
 ls -ld ~/.claude/plugins/marketplaces/specweave
-# Should show: lrwxr-xr-x ... -> /path/to/your/repo (SYMLINK)
-# NOT: drwxr-xr-x ... (DIRECTORY)
+# Must show: lrwxr-xr-x ... -> /path/to/repo (SYMLINK, not drwxr-xr-x)
 
-# If it's a directory, fix it:
+# If directory, fix it:
 rm -rf ~/.claude/plugins/marketplaces/specweave
 ln -s "$(pwd)" ~/.claude/plugins/marketplaces/specweave
 
-# Verify the fix:
+# Verify:
 bash .specweave/increments/0043-spec-md-desync-fix/scripts/verify-dev-setup.sh
 ```
 
-### Protection Measures
-
-1. ✅ **Pre-commit hook** verifies symlink before each commit (warns if broken)
-2. ✅ **Verification script** checks all aspects of setup (see "Local Development Setup" below)
-3. ✅ **This warning** ensures new contributors know about this issue
+**Without symlink**: Hooks fail, status line broken, living docs don't sync.
+**Protection**: Pre-commit hook verifies symlink before each commit.
 
-### Deep Dive
+### 2. NEVER Pollute Increment Folders
 
-For complete root cause analysis, investigation process, and prevention strategies:
-- **Ultrathink Report**: `.specweave/increments/0043-spec-md-desync-fix/reports/ULTRATHINK-HOOK-EXECUTION-ERRORS-ROOT-CAUSE-ANALYSIS-2025-11-18.md`
-- **Fixes Applied**: `.specweave/increments/0043-spec-md-desync-fix/reports/HOOK-EXECUTION-ERRORS-FIXES-APPLIED-2025-11-18.md`
+**ONLY 4 files allowed in `.specweave/increments/####/` root**:
+- `spec.md`, `plan.md`, `tasks.md`, `metadata.json`
 
-**Golden Rule**: If hooks fail with "No such file or directory", check the symlink FIRST!
+**Everything else → subfolders**:
+- `reports/` - Analysis, completion reports, validation
+- `scripts/` - Helper scripts, automation
+- `logs/` - Execution logs, debug output
 
----
-
-## 🚨 CRITICAL: NEVER POLLUTE PROJECT ROOT!
-
-**⛔ THIS IS THE #1 RULE - VIOLATING THIS WILL GET YOUR PR REJECTED ⛔**
-
-**ALL AI-generated files MUST go into the CURRENT INCREMENT folder**, NOT in the project root!
+**Examples**:
+```bash
+# ❌ WRONG
+.specweave/increments/0046/analysis-report.md
 
-### ❌ NEVER Create in Root (Pollutes Repository)
+# ✅ CORRECT
+.specweave/increments/0046/reports/analysis-report.md
 
+# Fix before committing:
+mkdir -p .specweave/increments/0046/reports
+mv .specweave/increments/0046/*.md .specweave/increments/0046/reports/
+git restore .specweave/increments/0046/{spec,plan,tasks}.md
 ```
-❌ WRONG - ROOT FILES (REJECTED!):
-/PLUGIN-MIGRATION-COMPLETE.md          # NO! Goes to increment reports/
-/SESSION-SUMMARY-2025-10-28.md         # NO! Goes to increment reports/
-/ADR-006-DEEP-ANALYSIS.md              # NO! Goes to .specweave/docs/internal/architecture/adr/
-/ANALYSIS-MULTI-TOOL-COMPARISON.md     # NO! Goes to increment reports/
-/QUICK-START.md                        # NO! Goes to increment reports/
-/migration-helper.sh                   # NO! Goes to increment scripts/
-/execution.log                         # NO! Goes to increment logs/
-
-✅ CORRECT - INCREMENT FOLDERS:
-.specweave/increments/0004-plugin-architecture/
-├── spec.md                            # ⚠️ ONLY THESE 3 FILES in root!
-├── plan.md
-├── tasks.md                           # Tasks with embedded tests
-├── reports/                           # ✅ ALL REPORTS HERE!
-│   ├── PLUGIN-MIGRATION-COMPLETE.md   # ✅ Completion reports
-│   ├── SESSION-SUMMARY.md             # ✅ Session summaries
-│   ├── QUICK-START.md                 # ✅ Quick start guides
-│   └── ANALYSIS-*.md                  # ✅ Analysis files
-├── scripts/                           # ✅ ALL SCRIPTS HERE!
-│   └── migration-helper.sh            # ✅ Helper scripts
-└── logs/                              # ✅ ALL LOGS HERE!
-    └── execution.log                  # ✅ Execution logs
-
-.specweave/docs/internal/architecture/ # ✅ PUT ADRS/DIAGRAMS HERE!
-└── adr/
-    └── 0006-deep-analysis.md          # ✅ Architecture decisions
-```
-
-**Before committing, ALWAYS check**: `git status` - If you see `.md` files in root, MOVE THEM!
-
-### 📁 Increment Structure Rules (MANDATORY)
 
-**ONLY 3 files allowed in increment root**:
-1. ✅ `spec.md` - Specification
-2. ✅ `plan.md` - Implementation plan
-3. ✅ `tasks.md` - Tasks with embedded tests
+### 3. NEVER Delete .specweave/ Directories
 
-**Everything else MUST be in subfolders**:
-- `reports/` - Session summaries, completion reports, analysis files, quick-start guides
-- `scripts/` - Helper scripts, migrations, utilities
-- `logs/` - Execution logs, debug output, temp files
+**Protected**: `.specweave/docs/`, `.specweave/increments/`
 
-**Examples of files that belong in subfolders**:
-- `QUICK-START.md` → `reports/QUICK-START.md`
-- `SESSION-NOTES.md` → `reports/SESSION-NOTES.md`
-- `ULTRATHINK-*.md` → `reports/ULTRATHINK-*.md`
-- `validation.sh` → `scripts/validation.sh`
-- `debug.log` → `logs/debug.log`
+**Pre-commit hook blocks**:
+- `rm -rf .specweave/docs` or `rm -rf .specweave/increments`
+- Deletion of 50+ files at once in these directories
 
-**Why this matters**:
-- ✅ Clean, predictable structure
-- ✅ Easy to find files by type
-- ✅ No increment root clutter
-- ✅ Consistent across all increments
+**If accidental deletion**:
+```bash
+git restore .specweave/
+```
 
----
+### 4. Test Cleanup Safety (MANDATORY)
 
-## 🛡️ CRITICAL: NEVER DELETE .specweave/ DIRECTORIES!
+**Danger**: `rm -rf` with wrong `pwd` = delete entire test suite.
 
-**⛔ MASS DELETION PROTECTION IS ACTIVE ⛔**
+**REQUIRED steps before any `rm -rf`**:
+1. Verify `pwd` (MUST be project root)
+2. Dry-run with `-print` (NO deletion)
+3. Count files/directories to delete
+4. Manual confirmation (type "DELETE")
+5. Execute deletion
+6. Verify results
+7. Run tests
 
-**PROTECTED DIRECTORIES**:
-- `.specweave/docs/` - All project documentation (internal + public)
-- `.specweave/increments/` - All increment history and specifications
+**Never** use `rm -rf` without ALL safety checks above.
 
-**WHAT THIS MEANS**:
-- ❌ **NEVER** run `rm -rf .specweave/docs` or `rm -rf .specweave/increments`
-- ❌ **NEVER** delete more than 50 files in these directories at once
-- ✅ Pre-commit hook will **BLOCK** accidental mass deletions
-- ✅ If intentional, bypass with `git commit --no-verify`
+### 5. NEVER Use `specweave init . --force`
 
-**WHY THIS EXISTS**:
-On 2025-11-17, an accidental mass deletion occurred (1,200+ files). All files were recovered via `git restore`, but this protection prevents future incidents.
+**Danger**: Deletes ALL increments and docs without backup (unless you bypass warnings).
 
-**IF YOU ACCIDENTALLY DELETE**:
+**Safe alternative**:
 ```bash
-# Immediately restore:
-git restore .specweave/
-
-# Verify restoration:
-git status
+specweave init .  # Interactive, never deletes without confirmation
 ```
 
-**See**: `.specweave/increments/0039/reports/ACCIDENTAL-DELETION-RECOVERY-2025-11-17.md`
+### 6. Increment Completion Validation
 
----
+**NEVER** mark increment "completed" without:
+1. All acceptance criteria checked (`- [x] **AC-...`)
+2. All tasks completed
+3. All tests passing
+4. Coverage target met
 
-## 🚨 CRITICAL: DANGEROUS TEST CLEANUP OPERATIONS!
+**Always use**: `/specweave:done 0043` (validates before closing)
+**Never**: Manual `metadata.json` edit (blocked by pre-commit hook)
 
-**⛔ WARNING: Test cleanup can cause CATASTROPHIC deletion ⛔**
+### 7. Source of Truth: tasks.md and spec.md (CRITICAL!)
 
-**INCIDENT**: On 2025-11-18, during increment 0042 (test infrastructure cleanup), a bash command nearly deleted the entire `tests/integration/` directory (209 test files). The command failed due to working directory confusion, but demonstrated the EXACT catastrophic risk identified in ultrathink analysis.
+**THE MOST CRITICAL RULE**: Internal TODO lists are ONLY for tracking work during execution. The SOURCE OF TRUTH is ALWAYS:
+1. **tasks.md** - Task completion status (`[x]` checkboxes)
+2. **spec.md** - Acceptance criteria status (`[x]` checkboxes)
 
-**THE DANGER**:
-```bash
-# ❌ EXTREMELY DANGEROUS (can delete entire test suite):
-cd tests/integration
-find . -maxdepth 1 -type d ... -exec rm -rf {} +
-# Risk: Working directory confusion + no confirmation = catastrophic deletion
-
-# ❌ DANGEROUS (no verification):
-find tests/integration -type d -exec rm -rf {} +
-# Risk: Too broad, deletes everything
-
-# ❌ DANGEROUS (assumes pwd):
-rm -rf tests/integration/*/
-# Risk: If pwd is wrong, deletes unintended directories
-```
+**MANDATORY Workflow When Using TodoWrite Tool**:
 
-**SAFE APPROACH** (MANDATORY for test cleanup):
-```bash
-# ✅ STEP 1: Verify working directory (ALWAYS FIRST!)
-pwd
-# Must output: /Users/antonabyzov/Projects/github/specweave
-
-# ✅ STEP 2: DRY RUN (list only, NO deletion)
-find tests/integration -maxdepth 1 -type d \
-  ! -name "integration" \
-  ! -name "core" ! -name "features" ! -name "external-tools" ! -name "generators" \
-  -print
-# Review output MANUALLY before proceeding
-
-# ✅ STEP 3: Count directories to delete
-find tests/integration -maxdepth 1 -type d \
-  ! -name "integration" \
-  ! -name "core" ! -name "features" ! -name "external-tools" ! -name "generators" \
-  -print | wc -l
-# Verify count matches expectations
-
-# ✅ STEP 4: Manual confirmation (REQUIRED)
-echo "About to delete XX directories. Type 'DELETE' to confirm:"
-read confirmation
-if [ "$confirmation" != "DELETE" ]; then
-  echo "Cancelled"
-  exit 1
-fi
-
-# ✅ STEP 5: Execute deletion (ONLY after all checks pass)
-find tests/integration -maxdepth 1 -type d \
-  ! -name "integration" \
-  ! -name "core" ! -name "features" ! -name "external-tools" ! -name "generators" \
-  -exec rm -rf {} +
-
-# ✅ STEP 6: Verify results
-ls tests/integration/
-find tests/integration -name "*.test.ts" | wc -l
-
-# ✅ STEP 7: Run tests (MANDATORY)
-npm run test:integration
 ```
-
-**MANDATORY SAFETY RULES**:
-1. ✅ **ALWAYS** verify `pwd` before ANY deletion command
-2. ✅ **ALWAYS** use dry-run (`-print`) before `-exec rm`
-3. ✅ **ALWAYS** count directories/files before deletion
-4. ✅ **ALWAYS** require manual confirmation for deletion
-5. ✅ **ALWAYS** verify results after deletion
-6. ✅ **ALWAYS** run tests after structural changes
-7. ✅ **NEVER** use `rm -rf` without multiple safety checks
-
-**WHY THIS MATTERS**:
-- 🔴 One wrong command = entire test suite deleted
-- 🔴 213 unsafe `process.cwd()` patterns exist in tests (increment 0042 audit)
-- 🔴 Working directory confusion is COMMON
-- 🔴 Recovery requires git restore (assumes uncommitted changes)
-- 🔴 If changes committed, recovery requires time-consuming rollback
-
-**LESSONS FROM INCIDENT 0042**:
-1. **Dry-run is NOT optional** - Always list before deleting
-2. **Working directory matters** - Verify pwd FIRST
-3. **Confirmation prevents accidents** - Require manual "DELETE" confirmation
-4. **Test after changes** - Structural changes can break tests
-5. **Incremental commits** - Commit after each phase (easier rollback)
-
-**IF CATASTROPHIC DELETION OCCURS**:
-```bash
-# Immediately restore from git (if not committed):
-git restore tests/
-
-# If committed, restore from previous commit:
-git log --oneline -- tests/
-git checkout <previous-commit-hash> -- tests/
-
-# Verify restoration:
-find tests -name "*.test.ts" | wc -l
-npm run test:all
+For EVERY task marked complete in internal TODO:
+1. ✅ Mark task as completed in internal TODO
+2. ⚠️  IMMEDIATELY update tasks.md checkbox: `[ ] pending` → `[x] completed`
+3. ⚠️  IMMEDIATELY update spec.md AC checkbox: `[ ]` → `[x]`
+4. ✅ Verify both files updated BEFORE moving to next task
 ```
 
-**REFERENCES**:
-- Incident report: `.specweave/increments/0042-test-infrastructure-cleanup/reports/CATASTROPHIC-DELETION-INCIDENT-2025-11-18.md`
-- Ultrathink analysis: `.specweave/increments/0041-living-docs-test-fixes/reports/ULTRATHINK-TEST-DUPLICATION-ANALYSIS-2025-11-18.md`
-- Safe cleanup approach: Increment 0042 Phase 1 tasks
+**Example of CORRECT workflow**:
+```typescript
+// Step 1: Complete the work
+await createIntegrationTest();
 
-**NEVER delete test directories without following ALL safety steps above!**
+// Step 2: Update internal TODO
+TodoWrite([{task: "T-013", status: "completed"}]);
 
----
+// Step 3: IMMEDIATELY update tasks.md (NEVER skip this!)
+Edit("tasks.md", "**Status**: [ ] pending", "**Status**: [x] completed");
 
-## ⚠️ CRITICAL: NEVER USE `specweave init . --force` FOR REINSTALLS!
+// Step 4: IMMEDIATELY update spec.md ACs
+Edit("spec.md", "- [ ] **AC-US1-01**", "- [x] **AC-US1-01**");
 
-**⛔ COMMON MISTAKE THAT DELETES ALL DATA ⛔**
-
-**THE DANGER**:
-```bash
-# ❌ DANGEROUS (deletes ALL increments and docs):
-specweave init . --force
-
-# What --force actually does:
-# 1. Skips all confirmation prompts
-# 2. AUTOMATICALLY DELETES .specweave/ entirely
-# 3. Loses all increments, docs, and history
-# 4. No backup unless you create one manually
+// Step 5: Move to next task
 ```
 
-**SAFE ALTERNATIVES**:
-```bash
-# ✅ SAFE - Update files, keep all data:
-specweave init .
-# When prompted, select: "Continue working"
-
-# ✅ SAFE - Always interactive, never deletes:
-npx specweave init .
+**❌ CRITICAL ERROR - Never Do This**:
+```typescript
+// ❌ WRONG: Marking TODO as complete WITHOUT updating source files
+TodoWrite([
+  {task: "T-013", status: "completed"},  // ❌ Only internal tracking
+  {task: "T-014", status: "completed"},  // ❌ Only internal tracking
+  {task: "T-015", status: "completed"}   // ❌ Only internal tracking
+]);
+// tasks.md still shows `[ ] pending` → DESYNC!
+// This is a CRITICAL VIOLATION!
 ```
 
-**WHY THIS MATTERS**:
-- Documentation used to recommend `--force` for troubleshooting (FIXED in v0.21.4+)
-- Users followed the docs and lost all their work
-- Now `--force` has multiple safeguards:
-  - ⚠️ BIG RED WARNING before deletion
-  - ✅ ALWAYS requires confirmation (even in force mode)
-  - 📦 Automatic backup created before deletion
-  - 🔒 Pre-commit hook blocks accidental commits
+**Pre-Closure Validation** (MANDATORY before `/specweave:done`):
+```bash
+# 1. Verify ALL tasks marked in tasks.md
+grep "^\*\*Status\*\*:" tasks.md | grep -c "\[x\] completed"
+# Output MUST equal total_tasks in frontmatter
 
-**IF YOU NEED A FRESH START**:
-1. Backup first: `cp -r .specweave .specweave.backup-$(date +%Y%m%d)`
-2. Run: `specweave init .` (select "Fresh start" option)
-3. Or: `specweave init . --force` (requires confirmation + creates auto-backup)
+# 2. Verify ALL ACs checked in spec.md
+grep -c "^- \[x\] \*\*AC-" spec.md
+# Output MUST equal total ACs
 
-**NEVER use `--force` unless you want to DELETE EVERYTHING!**
+# 3. Only then close increment
+/specweave:done 0044
+```
 
----
+**Why This Matters**:
+- Internal TODO is ephemeral (lost between sessions)
+- tasks.md is permanent source of truth
+- spec.md is contract with stakeholders
+- Desync = broken promises to users/team
 
-## Tool Support
+**Incident Reference**: 2025-11-19 - Increment 0044 was incorrectly closed with tasks.md showing `[ ] pending` while internal TODO showed "completed". This violated SpecWeave's core principle. See `.specweave/increments/0044-integration-testing-status-hooks/reports/INCIDENT-SOURCE-OF-TRUTH-VIOLATION.md` for full post-mortem.
 
-SpecWeave supports multiple AI coding tools with varying automation levels:
-- **Claude Code** (Recommended): Native support via plugins, hooks, MCP
-- **Cursor**: Partial support via AGENTS.md compilation
-- **Other tools** (Copilot, ChatGPT, Gemini): Basic support via AGENTS.md
+### 8. NEVER Use `console.*` in Production Code
 
-For adapter implementation details, see "Multi-Tool Support via Adapters" section below.
+**Rule**: ALL `src/` code MUST use logger abstraction, NEVER `console.log/error/warn`.
 
----
+**Why**: `console.*` pollutes test output even when errors are expected and properly handled.
 
-## Development Workflow
+**Use logger injection**:
+```typescript
+import { Logger, consoleLogger } from '../../utils/logger.js';
 
-### Core SpecWeave Commands
+export class MyClass {
+  private logger: Logger;
 
-**Note**: Detailed rules (naming, discipline, archiving) are in skills that auto-load when you use these commands.
+  constructor(options: { logger?: Logger } = {}) {
+    this.logger = options.logger ?? consoleLogger;
+  }
 
-**Primary Workflow**:
-```bash
-/specweave:increment "feature name"  # Plan new work (increment-planner skill)
-/specweave:do                        # Execute tasks
-/specweave:progress                  # Check status
-/specweave:done                      # Close increment (with validation)
+  doSomething() {
+    this.logger.log('Info message');
+    this.logger.error('Error message', error);
+  }
+}
 ```
 
-**🔥 CRITICAL: Increment Completion Validation**
-
-**NEVER** mark an increment as "completed" without ALL of the following:
-1. ✅ All acceptance criteria checked (`- [x] **AC-...`)
-2. ✅ All tasks completed (`**Status**: [x] completed`)
-3. ✅ All tests passing (unit, integration, E2E)
-4. ✅ Coverage target met (`npm run test:coverage`)
-
-**Automated Protection**:
-- `/specweave:done` runs `IncrementCompletionValidator` before PM validation
-- Pre-commit hook blocks commits with invalid completion status
-- CI/CD validates completion on pull requests
-
-**Manual Protection** (if editing metadata.json directly):
-```bash
-# ❌ NEVER DO THIS - Manual metadata edit without validation
-vim .specweave/increments/0043-spec-md-desync-fix/metadata.json
-# Change "status": "active" → "completed"
-git commit -m "Complete increment"  # Will be BLOCKED by pre-commit hook
+**In tests, use `silentLogger`**:
+```typescript
+import { silentLogger } from '../../src/utils/logger.js';
 
-# ✅ ALWAYS USE THIS - Let /specweave:done validate
-/specweave:done 0043  # Validates ACs, tasks, tests, docs before closing
+const instance = new MyClass({ logger: silentLogger });
 ```
 
-**What Happens if Validation Fails**:
+**Protection**: Code review + search before commit:
 ```bash
-/specweave:done 0043
-
-❌ CANNOT CLOSE INCREMENT - Automated validation failed
-
-  • 17 acceptance criteria still open
-  • 13 tasks still pending
-
-Fix these issues before running /specweave:done again
+# Check for console.* in src/ before committing:
+git diff --cached --name-only | grep '^src/.*\.ts$' | xargs grep -n 'console\.' 2>/dev/null
 ```
 
-**Why This Matters**:
-- **False completion** → Status line shows wrong increment
-- **GitHub sync broken** → Issues marked closed with open work
-- **Data integrity violation** → Source of truth is inconsistent
-- **Stakeholder confusion** → External tools show wrong status
+### 9. Coding Standards
 
-**State Management**:
-```bash
-/specweave:pause 0002 --reason="..." # Pause increment
-/specweave:resume 0002               # Resume paused
-/specweave:abandon 0002              # Abandon incomplete
-```
+**Full standards**: `.specweave/docs/internal/governance/coding-standards.md`
+**Auto-discovery**: Runs during brownfield analysis or `/specweave:analyze-standards`
 
-**Archiving** (MANUAL ONLY):
-```bash
-/specweave:archive 0031              # Archive specific
-/specweave:archive --keep-last 10    # Archive old work
-/specweave:restore 0031              # Restore from archive
-```
+**Critical rules (enforced during code generation)**:
 
-**Quality**:
-```bash
-/specweave:validate 0001             # Rule-based validation
-/specweave:qa 0001                   # AI quality assessment
-```
+1. ✅ **NEVER use `console.*`** - Use logger abstraction (already enforced above)
+2. ✅ **ALWAYS import with `.js` extensions** - Required for ESM (already enforced in Build section)
+3. ✅ **Test files MUST use `.test.ts`** suffix, never `.spec.ts` (already enforced in Testing section)
+4. **Avoid `any` type** - Use specific types or generics
+5. **Functions < 50 lines** (ideal), < 100 lines (max)
+6. **Use custom error types**, not generic Error
+7. **Comment "why" not "what"**
+8. **No hardcoded secrets** - Use environment variables
+9. **No N+1 queries** - Batch database operations
+10. **Naming**: camelCase (vars), PascalCase (classes), UPPER_SNAKE_CASE (constants)
 
-**Documentation**:
-```bash
-/specweave:sync-docs update          # Sync living docs
-/specweave:sync-tasks                # Sync task status
-```
+**Auto-discovery features**:
+- **Brownfield projects**: Standards auto-detected during project analysis
+- **Manual analysis**: `/specweave:analyze-standards` - Generate comprehensive standards report
+- **Drift detection**: `/specweave:analyze-standards --drift` - Check compliance with declared standards
+- **Update standards**: `/specweave:analyze-standards --update` - Update official standards from analysis
 
-**For complete command reference**: See "Quick Reference" section below.
+**How it works**:
+1. Scans codebase (src/**/*.ts) for naming patterns, import styles, function characteristics
+2. Parses ESLint, Prettier, TypeScript configs for enforced rules
+3. Analyzes existing CLAUDE.md, CONTRIBUTING.md for declared standards
+4. Generates evidence-based standards with confidence levels (90%+ = HIGH confidence)
+5. Detects anti-patterns: hardcoded secrets, large files (>500 lines), missing error handling
+6. Outputs to `.specweave/docs/internal/governance/coding-standards-analysis.md`
+
+**Note**: Most standards are enforced by ESLint/Prettier. This list focuses on SpecWeave-specific rules and patterns that can't be auto-fixed by linters.
 
 ---
 
 ## Project Structure
 
-### Source of Truth Principle
-
 ```
-src/                    # TypeScript code ONLY (compiled to dist/)
-plugins/                # ALL Claude Code components (skills, agents, commands, hooks)
-├── specweave/          # Core plugin (auto-loaded)
+src/                    # TypeScript code (compiled to dist/)
+plugins/                # Claude Code components (skills, agents, commands, hooks)
+├── specweave/          # Core plugin
 └── specweave-*/        # Optional plugins (GitHub, JIRA, etc.)
 .specweave/             # Framework data (increments, docs, logs)
 ```
 
-**Key Rules**:
-- ✅ `src/` = TypeScript code ONLY
-- ✅ ALL skills/agents/commands/hooks = Inside `plugins/`
-- ✅ Marketplace = GLOBAL via CLI (no per-project `.claude/`)
-- ❌ NEVER mix `*.ts` and `SKILL.md` in same directory
-- ❌ NEVER create new files in project root (use increment folders)
-
-**For complete structure**: See `README.md`
+**Rules**:
+- `src/` = TypeScript ONLY
+- ALL skills/agents/commands/hooks = `plugins/`
+- Marketplace = GLOBAL via CLI
+- NEVER mix `*.ts` and `SKILL.md` in same directory
+- NEVER create files in project root
 
 ---
 
-## Plugin Architecture
-
-### Core Plugin (Always Auto-Loaded)
-
-**Plugin**: `specweave` - The essential SpecWeave plugin loaded in every project:
-- **Skills**: 9 skills (increment-planner, tdd-workflow, spec-generator, etc.)
-- **Agents**: 22 agents (PM, Architect, Tech Lead, + 19 specialized)
-- **Commands**: 22 commands (/specweave:increment, /specweave:do, etc.)
-- **Hooks**: 8 lifecycle hooks
-- **Size**: ~12K tokens
-
-### Available Plugins
-
-All plugins are auto-installed during `specweave init`. Skills activate based on context keywords.
-
-**Plugin List**: `ls plugins/` or `/plugin list --installed`
-
-**Example plugins**:
-- `specweave` - Core (increment lifecycle, living docs)
-- `specweave-github` - GitHub Issues sync
-- `specweave-{frontend|backend|infrastructure}` - Tech stacks
-- `specweave-{ml|payments}` - Domain-specific
-
-### Plugin Installation
+## Development Workflow
 
-`specweave init` automatically:
-1. Registers marketplace: `claude plugin marketplace add anton-abyzov/specweave`
-2. Installs all plugins via Claude CLI
-3. Marketplace is GLOBAL (persists across projects)
+### Core Commands
 
-After init, all plugins are ready. Skills auto-activate based on keywords.
+```bash
+# Primary workflow
+/specweave:increment "feature"  # Plan new work
+/specweave:do                   # Execute tasks
+/specweave:progress             # Check status
+/specweave:done                 # Close (validates ACs, tasks, tests)
 
-### Local Development Setup (Contributors Only)
+# State management
+/specweave:pause 0002 --reason="..."
+/specweave:resume 0002
+/specweave:abandon 0002
 
-**🚨 CRITICAL for SpecWeave Contributors:**
+# Quality
+/specweave:validate 0001
+/specweave:qa 0001
 
-When developing SpecWeave itself, you MUST use a **symlink** from the marketplace to your local repository. This ensures:
-- ✅ All code/hook/skill changes are immediately reflected
-- ✅ No need to reinstall plugins after every change
-- ✅ Real-time testing of hooks, skills, and commands
+# Documentation
+/specweave:sync-docs update
+/specweave:sync-tasks
+```
 
-**⚡ QUICK SETUP (MANDATORY - Do this FIRST!):**
+### Local Development Setup
 
 ```bash
-# 1. Clone and install dependencies
+# 1. Clone and install
 git clone https://github.com/YOUR_USERNAME/specweave.git
 cd specweave
 npm install
 
-# 2. Create marketplace symlink (CRITICAL!)
-# ⚠️ IMPORTANT: Use absolute path, NOT $PWD (see ultrathink report for details)
+# 2. Create symlink (CRITICAL!)
 mkdir -p ~/.claude/plugins/marketplaces
-rm -rf ~/.claude/plugins/marketplaces/specweave  # Remove any existing directory/symlink
+rm -rf ~/.claude/plugins/marketplaces/specweave
 ln -s "$(pwd)" ~/.claude/plugins/marketplaces/specweave
 
-# 3. Verify setup (MANDATORY!)
+# 3. Verify setup
 bash .specweave/increments/0043-spec-md-desync-fix/scripts/verify-dev-setup.sh
-# Must show: ✅ ALL CHECKS PASSED!
-# If verification fails, see ultrathink report:
-#   .specweave/increments/0043-spec-md-desync-fix/reports/ULTRATHINK-HOOK-EXECUTION-ERRORS-ROOT-CAUSE-ANALYSIS-2025-11-18.md
 
-# 4. Install git hooks (for pre-commit verification)
+# 4. Install git hooks
 bash scripts/install-git-hooks.sh
 ```
 
-**Why Symlink is MANDATORY:**
-
-Claude Code's hook execution system looks for hooks in `~/.claude/plugins/marketplaces/specweave/`, NOT in your local repo.
-
-**⚠️ CRITICAL: Directory vs Symlink Issue**
-
-The marketplace path MUST be a **symlink**, NOT a **directory**:
-- ✅ **Symlink** (`lrwxr-xr-x`): All hook changes immediately reflected
-- ❌ **Directory** (`drwxr-xr-x`): Stale copy, hooks fail with "No such file or directory"
-
-**Without the symlink:**
-- ❌ All post-task-completion hooks fail with "No such file or directory"
-- ❌ Status line doesn't update automatically
-- ❌ Living docs don't sync after task completion
-- ❌ Increment metadata doesn't update
-- ❌ You'll waste hours debugging hook errors
-
-**Root Cause Analysis**: See comprehensive ultrathink report at:
-`.specweave/increments/0043-spec-md-desync-fix/reports/ULTRATHINK-HOOK-EXECUTION-ERRORS-ROOT-CAUSE-ANALYSIS-2025-11-18.md`
-
-**Detailed Setup (if Quick Setup fails):**
-
-```bash
-# 1. Remove any existing marketplace installation
-rm -rf ~/.claude/plugins/marketplaces/specweave
-
-# 2. Create symlink to your local SpecWeave repository
-ln -s /path/to/your/specweave/repo ~/.claude/plugins/marketplaces/specweave
-
-# Example:
-ln -s ~/Projects/github/specweave ~/.claude/plugins/marketplaces/specweave
-
-# 3. Verify symlink is correct
-readlink ~/.claude/plugins/marketplaces/specweave
-# Should output: /path/to/your/specweave/repo
-
-# 4. Verify hooks are accessible
-test -f ~/.claude/plugins/marketplaces/specweave/plugins/specweave/hooks/post-task-completion.sh && \
-  echo "✅ Hooks accessible" || echo "❌ Setup failed"
-
-# 5. Run automated verification
-bash .specweave/increments/0043-spec-md-desync-fix/scripts/verify-dev-setup.sh
-```
-
-**Verification Output:**
-
-```
-✅ ALL CHECKS PASSED! Local development setup is correct.
-You can now use TodoWrite and other tools without hook errors.
-```
-
-**If you see "Plugin not found" errors:**
-
-This means the plugin registry is out of sync. Fix it:
-
+**If "Plugin not found" errors**:
 ```bash
-# 1. Backup current registry
-cp ~/.claude/plugins/installed_plugins.json ~/.claude/plugins/installed_plugins.json.backup
-
-# 2. Clear registry (forces rediscovery)
 echo '{"version": 1, "plugins": {}}' > ~/.claude/plugins/installed_plugins.json
-
-# 3. Update marketplace
 claude plugin marketplace update specweave
-
-# 4. Restart Claude Code
-# All 25 plugins will be rediscovered from your local repo
-```
-
-**Verification:**
-
-```bash
-# Check symlink target
-readlink ~/.claude/plugins/marketplaces/specweave
-# Should output: /path/to/your/specweave/repo
-
-# Check all plugins are accessible
-ls ~/.claude/plugins/marketplaces/specweave/plugins/
-# Should list: specweave, specweave-github, specweave-jira, etc.
-```
-
-**Why This Matters:**
-
-Without the symlink, Claude Code will try to execute hooks from a non-existent location:
 ```
-❌ Plugin hook error: /bin/sh:
-   ~/.claude/plugins/marketplaces/specweave/plugins/specweave/hooks/user-prompt-submit.sh:
-   No such file or directory
-```
-
-**Troubleshooting:**
-
-If you see "Plugin not found in marketplace 'specweave'" errors:
-1. Check symlink exists: `ls -la ~/.claude/plugins/marketplaces/specweave`
-2. Verify it points to your repo: `readlink ~/.claude/plugins/marketplaces/specweave`
-3. Recreate symlink if needed (see setup instructions above)
-
-**What NOT to Do:**
-
-- ❌ Don't copy the repository - use a symlink
-- ❌ Don't use relative paths in symlink - use absolute paths
-- ❌ Don't register the marketplace via `claude plugin marketplace add` - symlink is enough for local dev
-
----
-
-## Multi-Tool Support via Adapters
-
-SpecWeave supports multiple AI coding tools through an adapter system. Tool selection happens during `specweave init`.
-
-**Adapter Architecture**:
-- **Location**: `src/adapters/` (interface, loader, tool-specific implementations)
-- **Selection**: Auto-detected or via `--adapter` flag
-- **Files**: Tool-specific files (`.cursorrules`, `AGENTS.md`, etc.)
-
-**Supported Tools**:
-
-| Tool | Automation Level | Implementation | Status |
-|------|------------------|----------------|--------|
-| **Claude Code** | Full | Native plugins (no adapter) | ✅ Recommended |
-| **Cursor** | Partial | AGENTS.md compilation | ✅ Supported |
-| **Generic** | Basic | AGENTS.md static file | ✅ Supported |
-
-**Key Differences**:
-- **Claude Code**: No adapter needed - uses native plugin system
-- **Cursor/Generic**: Require compilation step to generate AGENTS.md
-- **All tools**: Share same `.specweave/` data structure
-
-**For contributors**: Adapter code is in `src/adapters/`. Tests in `tests/unit/adapter-loader.test.ts`.
 
 ---
 
-## Development Principles
-
-See [.github/CONTRIBUTING.md](.github/CONTRIBUTING.md#core-development-principles) for:
-- Source of Truth Discipline
-- Documentation = Code
-- Testing Requirements
-- Incremental Development
-- Multi-Tool Support
-
----
-
-## Recent Architectural Enhancements (v0.18.3+)
-
-### Project-Specific Tasks in User Stories
-
-**New in v0.18.3**: User stories now include project-specific checkable task lists instead of just links to increment tasks.
-
-**Key Changes**:
-1. **TaskProjectSpecificGenerator** (`src/core/living-docs/task-project-specific-generator.ts`):
-   - Filters increment tasks by User Story ID (via AC-IDs)
-   - Optional project keyword filtering (backend vs frontend)
-   - Preserves completion status from increment tasks.md
-
-2. **User Story File Format** - New `## Tasks` section:
-```markdown
-## Tasks
+## Build & Test
 
-- [ ] **T-001**: Setup API endpoint
-- [x] **T-003**: Add DB migration (completed)
+### Build
 
-> **Note**: Tasks are project-specific. See increment tasks.md for full list
+```bash
+npm run rebuild    # Clean + build (use this during development)
+npm run build      # Compile TypeScript
+npm run clean      # Remove dist/
 ```
 
-3. **GitHub Sync** - Issues now have checkable task lists:
-   - Stakeholders can tick/untick tasks in GitHub
-   - Task completion syncs from user story files
-   - Backward compatible (falls back to legacy extraction)
-
-**Benefits**:
-- **Project Isolation**: Backend tasks ≠ Frontend tasks
-- **Traceability**: Each user story explicitly lists its tasks
-- **GitHub UX**: Checkable task lists in issues
-- **Completion Tracking**: Status preserved from increment
+**Critical**: Always import with `.js` extensions:
+```typescript
+// ✅ CORRECT
+import { foo } from './bar.js';
 
-**Data Flow**:
+// ❌ WRONG
+import { foo } from './bar';
 ```
-Increment tasks.md (All tasks, source of truth)
-    ↓
-TaskProjectSpecificGenerator (Filters by US + Project)
-    ↓
-User Story ## Tasks Section (Project-specific checkboxes)
-    ↓
-GitHub Issue (Checkable task list for stakeholders)
-```
-
-**See Also**:
-- Implementation: `.specweave/increments/0034-github-ac-checkboxes-fix/reports/PROJECT-SPECIFIC-TASKS-IMPLEMENTATION-COMPLETE.md`
-- Architecture: `.specweave/increments/0034-github-ac-checkboxes-fix/reports/ULTRATHINK-PROJECT-SPECIFIC-TASKS-ARCHITECTURE.md`
-
----
-
-## Build & Test
 
-### Build Commands
-
-```bash
-# ALWAYS use clean build during development (prevents TS5055 errors)
-npm run rebuild             # Clean + build (RECOMMENDED)
-
-# Or manual steps:
-npm run clean              # Remove dist/ folder
-npm run build              # Compile TypeScript
-```
+**Fix missing extensions**: `node scripts/fix-js-extensions.js`
 
 ### Testing
 
-**Test Framework**: **Vitest** (migrated from Jest on 2025-11-17)
+**Framework**: Vitest (migrated from Jest 2025-11-17)
 
 ```bash
-npm test                    # Smoke tests (quick validation)
-npm run test:unit           # Unit tests with Vitest
-npm run test:integration    # Integration tests with Vitest
+npm test                    # Smoke tests
+npm run test:unit           # Unit tests
+npm run test:integration    # Integration tests
 npm run test:e2e            # E2E tests (Playwright)
 npm run test:all            # All tests
 npm run test:coverage       # Coverage report
 ```
 
-**Why Vitest?**
-- ✅ ESM-native (no tsconfig hacks)
-- ✅ Faster than Jest
-- ✅ Better TypeScript integration
-- ✅ Native import.meta.url support
-- ✅ Modern, actively maintained
-
-**Test Organization** (4 categories):
-- `tests/unit/` - Pure logic tests (no I/O) - **Vitest**
-- `tests/plugin-validation/` - Plugin structure contracts
-- `tests/integration/` - 4 semantic categories - **Vitest**:
-  - `external-tools/` - GitHub, JIRA, ADO, Kafka sync
-  - `core/` - Core framework + workflows
-  - `generators/` - Code generation (frontend, backend, ML)
-  - `features/` - Feature plugins (Figma, i18n, diagrams, etc.)
-- `tests/e2e/` - Full user scenarios - **Playwright**
-
-**Test Naming Convention** (standardized 2025-11-18):
-- ✅ ALL tests use `.test.ts` extension (unit, integration, E2E)
-- ❌ NEVER use `.spec.ts` extension (deprecated)
-- **Why**: Consistent glob patterns, simpler test discovery, aligned with Vitest conventions
-
-**Writing Tests**:
-```typescript
-// Import from vitest (NOT jest)
-import { describe, it, expect, beforeEach, vi } from 'vitest';
-
-// Mocking
-vi.mock('fs/promises');
-const mockFn = vi.fn();
-vi.clearAllMocks();
-```
-
-**Details**: `.specweave/docs/internal/architecture/TEST-ORGANIZATION-PROPOSAL.md`
-
-### Test Isolation (CRITICAL - Prevents .specweave/ Deletion!)
-
-**🚨 MANDATORY FOR ALL TESTS creating .specweave/ structures:**
-
-**THE PROBLEM**: Tests using `process.cwd()` can accidentally delete the project `.specweave/` folder containing all your work!
-
-**CORRECT PATTERN** (ALWAYS use this):
-```typescript
-import * as os from 'os';
-import * as path from 'path';
-
-// ✅ SAFE: Uses isolated temp directory
-const testRoot = path.join(os.tmpdir(), 'test-name-' + Date.now());
-```
-
-**DANGEROUS PATTERN** (NEVER use this):
-```typescript
-// ❌ DANGER: Creates directories in project root!
-const testRoot = path.join(process.cwd(), '.test-something');
-const testPath = path.join(__dirname, '..', '.specweave', 'increments');
-```
+**Test structure**:
+- `tests/unit/` - Pure logic (no I/O)
+- `tests/integration/` - Organized by: `core/`, `external-tools/`, `generators/`, `features/`
+- `tests/e2e/` - Full user scenarios
 
-### Vitest Mock Best Practices
+**Naming**: ALL tests use `.test.ts` (NEVER `.spec.ts`)
 
-**🚨 MANDATORY FOR ALL TESTS using mocks:**
+### Writing Tests
 
-**CORRECT PATTERN** (vi.mocked() for type-safe mocks):
 ```typescript
+// ✅ CORRECT
 import { describe, it, expect, beforeEach, vi } from 'vitest';
 import fs from 'fs-extra';
 
-// Mock fs-extra BEFORE using it
 vi.mock('fs-extra');
 
-// Type-safe mocked functions
 const mockReadFile = vi.mocked(fs.readFile);
-const mockWriteFile = vi.mocked(fs.writeFile);
-const mockExistsSync = vi.mocked(fs.existsSync);
 
 beforeEach(() => {
   vi.clearAllMocks();
-  mockExistsSync.mockReturnValue(true);
   mockReadFile.mockResolvedValue('content');
 });
 ```
 
-**DANGEROUS PATTERN** (NEVER use this):
+**Critical anti-patterns**:
 ```typescript
-// ❌ WRONG: Invalid anyed<> syntax (pre-Vitest migration)
-const mockFs = fs as anyed<typeof fs> & {
-  readFile: anyedFunction<typeof fs.readFile>;
-};
-// This causes TypeScript errors and test failures!
-```
-
-**Why This Matters**:
-1. `vi.mocked()` provides full type safety
-2. Catches mock setup errors at compile time
-3. Works correctly with Vitest's mock system
-4. Prevents runtime "Cannot read properties of undefined" errors
-
-**Common Mock Gotchas**:
-- Always call `vi.clearAllMocks()` in beforeEach
-- Use `mockResolvedValue` for async functions
-- Use `mockReturnValue` for sync functions
-- Check that mocks are actually called: `expect(mockFn).toHaveBeenCalled()`
+// ❌ NEVER use jest APIs
+jest.fn()                    // Use: vi.fn()
+jest.mock()                  // Use: vi.mock()
 
-### Testing Best Practices & Anti-Patterns
-
-**Lessons from fixing 72+ test failures (2025-11-17 Vitest migration)**
-
-#### Critical Rules (Will Fail Tests)
+// ❌ NEVER use project root in tests
+const testRoot = path.join(process.cwd(), '.test-something');
 
-```typescript
-// ❌ WRONG                           // ✅ CORRECT
-require('../src/module')              import { Module } from '../src/module.js'
-jest.fn()                             vi.fn()
-jest.mock()                           vi.mock()
-fs as anyed<typeof fs>                vi.mocked(fs.readFile)
-const mock = vi.fn()                  vi.mock('mod', () => ({
-vi.mock('mod', () => ({ mock }))        method: vi.fn()  // inline!
-                                      }))
-mockReadJson.mockResolvedValue(arr)   mockReadJson.mockResolvedValue([...arr])  // copy!
+// ✅ ALWAYS use temp directory
+const testRoot = path.join(os.tmpdir(), 'test-' + Date.now());
 ```
 
-#### Quick Template
-
+**Use test utilities**:
 ```typescript
-import { describe, it, expect, vi } from 'vitest';
-vi.mock('fs-extra', () => ({ default: { readFile: vi.fn() } }));
-import fs from 'fs-extra';
-
-describe('Feature', () => {
-  const mockReadFile = vi.mocked(fs.readFile);
-
-  beforeEach(() => {
-    vi.clearAllMocks();
-    mockReadFile.mockResolvedValue('data');
-  });
+import { createIsolatedTestDir } from '../test-utils/isolated-test-dir';
 
-  it('works', async () => {
-    const result = await MyModule.doSomething();
-    expect(mockReadFile).toHaveBeenCalled();
-  });
-});
+const { testDir, cleanup } = await createIsolatedTestDir('my-test');
+try {
+  // Test code here
+} finally {
+  await cleanup();  // ALWAYS cleanup
+}
 ```
 
-#### Pre-Commit Checklist
-
-- [ ] ES6 imports (NOT require())
-- [ ] vi.* APIs (NOT jest.*)
-- [ ] vi.mocked() for mocks (NOT anyed<>)
-- [ ] Inline mock factories (NO external variables)
-- [ ] Array copies in mocks (NO shared references)
-- [ ] vi.clearAllMocks() in beforeEach()
-
-**Full guide**: `tests/test-template.test.ts`
+### Build Architecture
 
-**Why This Matters**:
-1. Tests create mock `.specweave/` structures for testing
-2. Cleanup uses `fs.rm(testRoot, { recursive: true })`
-3. If `testRoot` points to project root → **DELETES REAL .specweave/!**
-4. You lose all increments, docs, and history
+**Dual compilation**:
+- `tsc`: `src/` → `dist/src/` (source files)
+- `esbuild`: `plugins/**/lib/hooks/*.ts` → in-place `.js` (hooks only)
 
-**Use Test Utilities** (RECOMMENDED):
+**Hook imports**:
 ```typescript
-import { createIsolatedTestDir, createSpecweaveStructure } from '../test-utils/isolated-test-dir';
-
-test('my test', async () => {
-  const { testDir, cleanup } = await createIsolatedTestDir('my-test');
-
-  try {
-    // Setup .specweave structure in isolated directory
-    await createSpecweaveStructure(testDir);
-
-    // Test code here - NEVER touches project .specweave/
-    const incrementPath = path.join(testDir, '.specweave', 'increments', '0001-test');
-    // ...
-  } finally {
-    await cleanup(); // ALWAYS cleanup
-  }
-});
-```
-
-**Protection Layers**:
-1. ✅ **Pre-commit hook**: Blocks commits with dangerous test patterns
-2. ✅ **Test utilities**: `tests/test-utils/isolated-test-dir.ts`
-3. ✅ **Documentation**: This section
-
-**Related Incident**: 2025-11-17 - Multiple `.specweave/` deletions traced to dangerous test patterns
-**Root Cause Analysis**: `.specweave/increments/0037/reports/DELETION-ROOT-CAUSE-2025-11-17.md`
-
-### Build Health Checks
-
-**CRITICAL**: TypeScript ES Modules require specific practices:
-
-1. **Always Import with .js Extensions**:
-   ```typescript
-   // ❌ WRONG (will break at runtime):
-   import { foo } from './bar';
-
-   // ✅ CORRECT:
-   import { foo } from './bar.js';
-   ```
-
-2. **Fix Missing Extensions**:
-   ```bash
-   # Auto-fix all missing .js extensions:
-   node scripts/fix-js-extensions.js
-
-   # Preview changes first:
-   node scripts/fix-js-extensions.js --dry-run
-   ```
-
-3. **Verify Build is Clean**:
-   ```bash
-   # Check for polluted dist/ (TS5055 indicator):
-   find dist/src -name "*.ts" -not -name "*.d.ts"
-   # Should return NOTHING
-
-   # If files found, clean rebuild:
-   npm run rebuild
-   ```
-
-4. **Test Hook Execution**:
-   ```bash
-   # Hooks must import correctly at runtime:
-   node plugins/specweave/lib/hooks/update-ac-status.js 0001
-   # Should execute without "Cannot find module" errors
-   ```
-
-5. **Install Git Hooks** (RECOMMENDED):
-   ```bash
-   bash scripts/install-git-hooks.sh
-   # Verifies build health before every commit
-   ```
-
-### Common Build Errors & Fixes
-
-**Error: TS5055 - Cannot write file (would overwrite input)**
-```bash
-# Cause: dist/ polluted with .ts source files
-# Fix:
-npm run clean && npm run build
-```
-
-**Error: Cannot find module 'src/...' (hook execution)**
-```bash
-# Cause: Hooks importing from src/ instead of dist/src/
-#    OR: Missing .js extensions in imports
-# Fix:
-node scripts/fix-js-extensions.js
-npm run rebuild
-```
-
-**Error: Module import without .js extension**
-```bash
-# Cause: TypeScript ES modules REQUIRE .js in relative imports
-# Fix:
-node scripts/fix-js-extensions.js
+#!/usr/bin/env node
+import { ACStatusManager } from '../../../../dist/src/core/...';  // Use dist/
 ```
 
-### Build Architecture
-
-SpecWeave uses **dual compilation**:
-- `tsc`: Compiles `src/` → `dist/src/` (source files)
-- `esbuild`: Compiles `plugins/**/lib/hooks/*.ts` → in-place `.js` (hooks only)
-
-**Why?** Hooks must import from `dist/src/` (compiled), but TypeScript would try to compile them before `dist/` exists (chicken-and-egg). Solution: Exclude hooks from `tsconfig.json`, compile separately with esbuild.
-
-### Increment Scripts vs Hooks
-
-**CRITICAL DISTINCTION**: Increment scripts and hooks have different purposes and import patterns.
-
-**Hooks** (`plugins/**/lib/hooks/*.ts`):
-- **Purpose**: Production runtime executables (called by shell scripts)
-- **Execution**: `#!/usr/bin/env node` (JavaScript runtime)
-- **Imports**: MUST use `dist/src/` (compiled JavaScript)
-- **Compilation**: esbuild (separate from main build)
-- **Example**:
-  ```typescript
-  #!/usr/bin/env node
-  import { ACStatusManager } from '../../../../dist/src/core/increment/ac-status-manager.js';
-  ```
-
-**Increment Scripts** (`.specweave/increments/####/scripts/*.ts`):
-- **Purpose**: Development utilities (one-off tasks, migrations, analysis)
-- **Execution**: `#!/usr/bin/env tsx` (TypeScript runtime with ESM support)
-- **Imports**: MUST use `src/` (TypeScript source, NOT compiled)
-- **Compilation**: None (tsx transpiles on-the-fly)
-- **Example**:
-  ```typescript
-  #!/usr/bin/env tsx
-  import { ACStatusManager } from '../../../../src/core/increment/ac-status-manager.js';
-  ```
-
-**Why Different Patterns?**
-
-| Aspect | Hooks (dist/) | Increment Scripts (src/) |
-|--------|---------------|-------------------------|
-| **Context** | Production runtime | Development only |
-| **Tooling** | Node.js only | TypeScript tooling (tsx) |
-| **Build dependency** | Yes (requires dist/) | No (live source code) |
-| **Live reload** | No (must rebuild) | Yes (tsx auto-transpiles) |
-| **Purpose** | Permanent infrastructure | Temporary dev utilities |
-
-**Common Mistake**: Copying hook import patterns to increment scripts.
-
+**Increment script imports**:
 ```typescript
-// ❌ WRONG (increment script importing from dist/):
-#!/usr/bin/env tsx
-import { ACStatusManager } from '../../../../dist/src/core/...';
-// Issues:
-// - Requires npm run build before execution
-// - Stale code if src/ changes without rebuild
-// - Semantic confusion (TS runtime + JS imports)
-
-// ✅ CORRECT (increment script importing from src/):
 #!/usr/bin/env tsx
-import { ACStatusManager } from '../../../../src/core/...';
-// Benefits:
-// - No build dependency
-// - Always uses latest source code
-// - Clear dev script pattern
+import { ACStatusManager } from '../../../../src/core/...';  // Use src/
 ```
 
-**Note**: Use `tsx` (not `ts-node`) for increment scripts - it has better ESM + TypeScript support.
-
-### Running Increment Scripts
-
-**CRITICAL**: Increment scripts use `#!/usr/bin/env tsx` shebang, which assumes tsx is in your PATH. This often fails in development environments.
-
-**Recommended Execution Method** (Always Works):
+**Running increment scripts**:
 ```bash
-# ✅ Use npx - finds tsx in node_modules/.bin/ automatically
 npx tsx .specweave/increments/####/scripts/script-name.ts
-
-# Example:
-npx tsx .specweave/increments/0037-project-specific-tasks/scripts/validate-task-consistency.ts
-```
-
-**Why npx?**
-- ✅ Works without global tsx installation
-- ✅ Uses project's tsx version (consistent across team)
-- ✅ No PATH configuration needed
-- ✅ Same behavior in CI/CD and local dev
-
-**Alternative Methods** (Less Recommended):
-```bash
-# Option 1: Direct execution (requires global tsx)
-chmod +x script.ts
-./script.ts
-# ❌ Fails if tsx not globally installed
-
-# Option 2: Install tsx globally (not recommended for teams)
-npm install -g tsx
-./script.ts
-# ❌ Every contributor must install globally
-# ❌ Version inconsistencies across team
-
-# Option 3: Full path
-./node_modules/.bin/tsx script.ts
-# ✅ Works, but verbose
-```
-
-**Troubleshooting "command not found: tsx" (Exit Code 127)**:
-
-This error means tsx isn't in your shell's PATH.
-
-**Root Cause**:
-- tsx is a **dev dependency** (installed in `node_modules/`)
-- Shell can't find commands in `node_modules/.bin/` automatically
-- npm scripts add `node_modules/.bin/` to PATH, but direct shell execution doesn't
-
-**Fix**:
-```bash
-# Instead of:
-tsx script.ts          # ❌ Fails with error 127
-
-# Use:
-npx tsx script.ts      # ✅ Always works
 ```
 
-**Why This Happens "Sometimes"**:
-- **npm scripts**: Work (PATH includes node_modules/.bin/)
-- **Direct shell**: Fails (PATH doesn't include node_modules/.bin/)
-- **Global tsx installed**: Works (tsx in system PATH)
-- **CI/CD**: Depends on environment setup
-
-**Execution Context Comparison**:
-
-| Method | tsx in PATH? | Works? |
-|--------|--------------|--------|
-| `npm run script` | ✅ Yes (npm adds it) | ✅ Yes |
-| `./script.ts` | ❌ No (unless global) | ❌ Usually fails |
-| `tsx script.ts` | ❌ No (unless global) | ❌ Usually fails |
-| `npx tsx script.ts` | ✅ Yes (npx finds it) | ✅ Always works |
-
-**Best Practice for Contributors**:
-- **Creating scripts**: Use `#!/usr/bin/env tsx` shebang (conventional)
-- **Running scripts**: Always use `npx tsx script.ts` (reliable)
-- **Documentation**: Show npx execution method in examples
-
-### Coverage Requirements
+### Coverage
 
 - Critical paths: 90%+
 - Overall: 80%+
 
-### Related Documentation
-
-- Build process details: `.github/CONTRIBUTING.md` → "Build Process & Best Practices"
-- Ultrathink analysis: `.specweave/increments/0039/reports/HOOK-IMPORT-ERROR-ULTRATHINK-ANALYSIS.md`
-- Build verification tests: `tests/integration/build/build-verification.test.ts`
-
 ---
 
 ## Common Tasks
 
-### Adding Skills, Agents, or Commands
+### Adding Components
 
-**All components go into plugins**:
-- **Core components**: `plugins/specweave/{skills|agents|commands|hooks}/`
-- **Plugin components**: `plugins/specweave-{name}/{skills|agents|commands}/`
-- **Tests**: `tests/integration/{component-name}/` or `tests/unit/`
+All components go into `plugins/`:
+- Core: `plugins/specweave/{skills|agents|commands|hooks}/`
+- Plugins: `plugins/specweave-{name}/{skills|agents|commands}/`
+- Tests: `tests/integration/` or `tests/unit/`
 
-**See**: `.github/CONTRIBUTING.md` for complete instructions
+See `.github/CONTRIBUTING.md` for details.
 
 ### Updating Documentation
 
 ```bash
-# Internal docs (architecture, ADRs)
+# Internal docs
 vim .specweave/docs/internal/architecture/hld-system.md
 
-# Public docs (user-facing guides)
+# Public docs
 vim .specweave/docs/public/guides/user-guide.md
-
-# Build docs site
-cd docs-site && npm run build
 ```
 
 ---
 
 ## Troubleshooting
 
-**Skills not activating?**
-1. Check plugin installed: `/plugin list --installed`
-2. Verify YAML frontmatter in skill SKILL.md
-3. Restart Claude Code
-4. Check description has trigger keywords
-
-**Commands not working?**
-1. Check plugin installed: `/plugin list --installed`
-2. Verify command exists in plugin
-3. Restart Claude Code
-
-**Tests failing?**
-1. Run `npm run build` first
-2. Check test output
-3. Verify test data in `tests/fixtures/`
-
-**Root folder polluted?**
-1. Move files to `.specweave/increments/####/reports/`
-2. Update `.gitignore` if needed
-
-**Plugin hooks not working? (Development)**
-See `.claude/CONTRIBUTING.md` for plugin marketplace setup.
-
----
-
-## Getting Help
-
-**Documentation**:
-- User docs: https://spec-weave.com
-- Contributor docs: `.specweave/docs/internal/`
-- Architecture: `.specweave/docs/internal/architecture/`
-
-**Community**:
-- GitHub Issues: https://github.com/anton-abyzov/specweave/issues
-- Discussions: https://github.com/anton-abyzov/specweave/discussions
-
-**Current Increment**:
-```bash
-/specweave:status  # Show all increments
-```
+**Skills not activating**: Check YAML frontmatter, restart Claude Code
+**Commands not working**: Verify plugin installed, restart Claude Code
+**Tests failing**: Run `npm run rebuild`, check test output
+**Root folder polluted**: Move files to `.specweave/increments/####/reports/`
+**Hooks failing**: Verify symlink (see "Symlink Setup" above)
 
 ---
 
 ## Quick Reference
 
-**Commands** (all use `/specweave:*` namespace):
-
-*Primary*:
-- `/specweave:increment "feature"` - Plan new increment
-- `/specweave:do` - Execute tasks
-- `/specweave:done 0002` - Close increment
-- `/specweave:validate 0002` - Validate increment
-- `/specweave:qa 0002` - Quality assessment
+**Commands**:
+- `/specweave:increment "feature"` - Plan
+- `/specweave:do` - Execute
+- `/specweave:done 0002` - Close (validates)
 - `/specweave:status` - Show status
 - `/specweave:progress` - Check progress
-
-*State management*:
-- `/specweave:pause 0002 --reason="..."` - Pause
-- `/specweave:resume 0002` - Resume
-- `/specweave:abandon 0002 --reason="..."` - Abandon
-
-*Archiving (MANUAL ONLY)*:
-- `/specweave:archive 0031` - Archive specific
-- `/specweave:archive --keep-last 10` - Archive old
-- `/specweave:restore 0031` - Restore
-- `/specweave:fix-duplicates` - Resolve duplicates
-
-*Documentation*:
+- `/specweave:validate 0002` - Validate
+- `/specweave:qa 0002` - Quality check
+- `/specweave:pause/resume/abandon` - State management
+- `/specweave:archive/restore` - Archiving (manual only)
 - `/specweave:sync-docs update` - Sync living docs
-- `/specweave:sync-tasks` - Sync task status
 
-*Other*:
-- `/specweave:costs` - AI cost dashboard
-- `/specweave:translate` - Translate content
-- `/specweave:update-scope` - Log scope changes
-- `/specweave:next` - Smart transition
-
-**Build & Test**:
-- `npm run build` - Compile TypeScript
-- `npm test` - Unit tests
-- `npm run test:e2e` - E2E tests
-- `npm run test:integration` - Integration tests
+**Build**:
+- `npm run rebuild` - Clean + build
+- `npm test` - Smoke tests
+- `npm run test:all` - All tests
+- `npm run test:coverage` - Coverage
 
 **File Structure**:
-- Source: `src/` (TypeScript) and `plugins/` (skills/agents/commands)
-- Marketplace: GLOBAL (via CLI)
+- Source: `src/` (TypeScript), `plugins/` (skills/agents/commands/hooks)
 - Increments: `.specweave/increments/`
-- Docs: `.specweave/docs/internal/` and `.specweave/docs/public/`
+- Docs: `.specweave/docs/internal/`, `.specweave/docs/public/`
 - Tests: `tests/` (unit, integration, E2E)
 
----
-
 **Remember**:
-1. Edit source files in `src/` and `plugins/`, not `.claude/`
-2. Keep root folder clean (use increment folders)
-3. Test before committing (E2E + unit + integration)
-4. Skills/agents/commands auto-activate when needed
-5. All detailed rules are in skills (progressive disclosure)
-
-**SpecWeave Documentation**: https://spec-weave.com
-**Last Updated**: 2025-11-15
+1. Symlink setup is MANDATORY for contributors
+2. Keep root clean (use increment folders)
+3. Test before committing
+4. Never delete .specweave/ directories
+5. Use `/specweave:done` (never manual metadata edits)
+
+**See also**: `.github/CONTRIBUTING.md`, https://spec-weave.com