feat: Move specs and memory to git for multi-developer collaboration (v1.0.6)

BREAKING CHANGE: Specs, plans, and memory moved from .auto-build-data/ (gitignored)
to .auto-build/ (git-tracked) to enable team collaboration.

Changes:
- Specs/plans now in .auto-build/specs/ (shared with team)
- Memory (patterns, gotchas) now in .auto-build/memory/ (shared with team)
- .auto-build-data/ now only contains local data (worktrees, cache)
- Added /ab:migrate command for existing projects
- Removed symlinks from worktree-create.sh (no longer needed)

Benefits:
- Any developer can continue a plan started by another
- Patterns and gotchas shared across team
- Works on Windows/Linux/Mac without symlinks
- Full version history in git

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

.claude-plugin/marketplace.json

@@ -11,7 +11,7 @@
       "name": "ab",
       "source": "./plugin",
       "description": "Spec-driven build orchestration with git worktree isolation and session memory",
-      "version": "1.0.5",
+      "version": "1.0.6",
       "keywords": ["build", "spec", "automation", "worktree", "qa"]
     }
   ]

CHANGELOG.md

@@ -2,6 +2,27 @@
 
 All notable changes to Auto-Build plugin.
 
+## [1.0.6] - 2024-12-22
+
+### Changed (BREAKING)
+- **Specs and plans moved to `.auto-build/`** (git-tracked) for multi-developer collaboration
+- **Memory moved to `.auto-build/memory/`** (git-tracked) so team shares learned patterns
+- `.auto-build-data/` now only contains local data (worktrees, cache)
+
+### Added
+- New `/ab:migrate` command for existing projects to migrate from old structure
+- Documentation updates explaining git-tracked vs local data separation
+
+### Removed
+- Symlinks in worktree-create.sh - no longer needed since data is in git
+- Old `.auto-build-data/specs/` and `.auto-build-data/memory/` paths
+
+### Benefits
+- 🤝 Multi-developer collaboration - any developer can continue a plan
+- 🧠 Shared learning - patterns and gotchas available to entire team
+- 🌍 Cross-platform - works on Windows/Linux/Mac without symlinks
+- 📜 Version history - specs/plans/memory changes visible in git
+
 ## [1.0.5] - 2024-12-22
 
 ### Fixed

README.md

@@ -45,7 +45,8 @@ Auto-Build este un plugin oficial Claude Code care transformă modul în care de
 #   - CLAUDE.md (dacă nu există)
 #   - .claude/rules/auto-build-patterns.md
 #   - .claude/rules/auto-build-memory.md
-#   - .auto-build-data/ structure
+#   - .auto-build/ structure (specs, memory - git-tracked)
+#   - .auto-build-data/ structure (worktrees, cache - local only)
 
 # 4. Verifică instalarea
 /ab:help
@@ -227,9 +228,9 @@ What insights should be saved?
 
 Agent salvează în **2 locații simultan**:
 
-**A) JSON Files** (searchable)
+**A) JSON Files** (searchable, git-tracked - shared with team)
 ```json
-// .auto-build-data/memory/patterns.json
+// .auto-build/memory/patterns.json
 {
   "id": "pat_20250121_143000",
   "title": "API Error Handling Pattern",
@@ -512,20 +513,28 @@ user-project/
 │   │   └── auto-build-memory.md    ← Updated de /ab:memory-save
 │   └── settings.json
 │
-└── .auto-build-data/               ← Runtime data (gitignored)
+├── .auto-build/                    ← COMMITTED TO GIT - shared with team
-    ├── specs/{feature-name}/
+│   ├── specs/{feature-name}/
-    │   ├── spec.md
+│   │   ├── spec.md
-    │   ├── plan.md
+│   │   ├── plan.md
-    │   └── status.json
+│   │   └── status.json
-    ├── worktrees/
+│   └── memory/
-    │   └── worktree-registry.json
-    ├── memory/
 │       ├── patterns.json           ← Searchable via /ab:memory-search
 │       ├── gotchas.json            ← Searchable via /ab:memory-search
 │       └── sessions/
+│
+└── .auto-build-data/               ← GITIGNORED - local only
+    ├── worktrees/
+    │   └── worktree-registry.json
     └── cache/qa-iterations/
 ```
 
+**Beneficii noi (v1.0.6)**:
+- 🤝 **Colaborare**: Alt developer poate continua planul tău
+- 🧠 **Învățare partajată**: Patterns și gotchas disponibile pentru toată echipa
+- 🌍 **Cross-platform**: Funcționează fără symlinks pe Windows/Linux/Mac
+- 📜 **Versionare**: Istoricul specs/plans/memory vizibil în git
+
 ---
 
 ## 🔧 Configurare Gitea

plugin/.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "ab",
   "description": "Auto-Build: Spec-driven build orchestration with worktree isolation and session memory",
-  "version": "1.0.5",
+  "version": "1.0.6",
   "author": {
     "name": "ROA2WEB Team"
   }

plugin/README.md

@@ -31,7 +31,7 @@ Auto-Build is a Claude Code plugin that provides autonomous feature implementati
        ↓
    spec-writer agent
        ↓
-   spec.md created in .auto-build-data/specs/
+   spec.md created in .auto-build/specs/ (git-tracked!)
        ↓
 /ab:build feature-name
        ↓
@@ -49,10 +49,12 @@ Auto-Build is a Claude Code plugin that provides autonomous feature implementati
        ↓
 /ab:memory-save
        ↓
-   patterns.json, gotchas.json updated
+   patterns.json, gotchas.json in .auto-build/memory/ (git-tracked!)
    + .claude/rules/auto-build-memory.md synced (auto-loaded!)
 ```
 
+**New in v1.0.6**: Specs, plans, and memory are now git-tracked, enabling multi-developer collaboration!
+
 ---
 
 ## Commands
@@ -130,20 +132,26 @@ user-project/                       # Your project
 │   │   └── auto-build-memory.md    # Updated by /ab:memory-save
 │   └── settings.json
 │
-└── .auto-build-data/               # Runtime data (gitignored)
+├── .auto-build/                    # ⭐ COMMITTED TO GIT - shared with team
-    ├── specs/{feature-name}/
+│   ├── specs/{feature-name}/
-    │   ├── spec.md
+│   │   ├── spec.md                 # Feature specification
-    │   ├── plan.md
+│   │   ├── plan.md                 # Implementation plan
-    │   └── status.json
+│   │   └── status.json             # Build status
-    ├── worktrees/
+│   └── memory/
-    │   └── worktree-registry.json
-    ├── memory/
 │       ├── patterns.json           # Searchable via /ab:memory-search
 │       ├── gotchas.json            # Searchable via /ab:memory-search
-    │   └── sessions/
+│       └── sessions/               # Session history
-    └── cache/qa-iterations/
+│
+└── .auto-build-data/               # GITIGNORED - local only
+    ├── worktrees/
+    │   └── worktree-registry.json  # Local worktree tracking
+    └── cache/qa-iterations/        # Temporary QA cache
 ```
 
+**Data Separation (v1.0.6)**:
+- `.auto-build/` - Git-tracked: specs, plans, memory (shared with team)
+- `.auto-build-data/` - Gitignored: worktrees, cache (local only)
+
 ---
 
 ## Agents
@@ -236,9 +244,9 @@ Auto-Build is designed to be portable:
 1. **Plugin-based**: Installs via official Claude Code plugin system
 2. **Self-contained**: Everything lives in plugin directory
 3. **No external dependencies**: Uses only Claude Code agents and Bash scripts
-4. **Clean separation**: Runtime data in `.auto-build-data/` (gitignored)
+4. **Clean separation**: Shared data in `.auto-build/` (git), local data in `.auto-build-data/` (gitignored)
 5. **Cross-platform**: Uses @import (git-friendly), not symlinks
-6. **Team sharing**: Git-friendly, works on all platforms
+6. **Team sharing**: Specs, plans, and memory are git-tracked for multi-developer collaboration
 
 ---
 

plugin/commands/build.md

@@ -13,15 +13,15 @@ Orchestrate the complete implementation of a feature using its specification. Th
 
 ## Prerequisites
 
-- Spec must exist at `.auto-build-data/specs/{feature-name}/spec.md`
+- Spec must exist at `.auto-build/specs/{feature-name}/spec.md`
 - Status must be `SPEC_COMPLETE` or `PLANNING_COMPLETE` or `IMPLEMENTING`
 
 ## Workflow
 
 ### 1. Load Specification
 
-- Read `.auto-build-data/specs/{feature-name}/spec.md`
+- Read `.auto-build/specs/{feature-name}/spec.md`
-- Read `.auto-build-data/specs/{feature-name}/status.json`
+- Read `.auto-build/specs/{feature-name}/status.json`
 - Validate spec exists and is complete
 - If status is `IMPLEMENTING`, resume from current task
 
@@ -37,18 +37,17 @@ Create an isolated git worktree? (recommended for larger features)
 If yes:
 - Run: `bash ${CLAUDE_PLUGIN_ROOT}/scripts/worktree-create.sh {feature-name}`
 - Update status.json with worktree path
-- **IMPORTANT**: After worktree creation and planning, display clear next steps:
+- Display next steps:
 
 ```
 ════════════════════════════════════════════════════════════════
-  WORKTREE CREATED - ACTION REQUIRED
+  WORKTREE CREATED
 ════════════════════════════════════════════════════════════════
 
 ✅ Worktree created at: ../ab-worktrees/{project}-{feature-name}/
 ✅ Branch: feature/ab-{feature-name}
-✅ Plan created: .auto-build-data/specs/{feature-name}/plan.md
 
-⚠️  NEXT STEPS (you must do this to continue):
+⚠️  NEXT STEPS:
 
 1. Open a NEW VSCode window in the worktree:
 
@@ -58,11 +57,9 @@ If yes:
 
    /ab:build {feature-name}
 
-   (It will detect the plan and start implementation)
-
 ────────────────────────────────────────────────────────────────
-Why? Claude Code sessions are tied to the current directory.
+Note: Specs and plans in .auto-build/ are already available in
-To work in the isolated worktree, you need a new session there.
+the worktree (they're in git). No symlinks needed!
 ════════════════════════════════════════════════════════════════
 ```
 
@@ -78,15 +75,15 @@ To work in the isolated worktree, you need a new session there.
 **If status is `SPEC_COMPLETE`**:
 
 1. Load memory context (if exists):
-   - Read `.auto-build-data/memory/patterns.json`
+   - Read `.auto-build/memory/patterns.json`
-   - Read `.auto-build-data/memory/gotchas.json`
+   - Read `.auto-build/memory/gotchas.json`
 
 2. Launch **planner** agent with:
    - The complete spec.md content
    - Relevant patterns from memory
    - Instruction to create ordered implementation tasks
 
-3. The planner creates `.auto-build-data/specs/{feature-name}/plan.md`:
+3. The planner creates `.auto-build/specs/{feature-name}/plan.md`:
    ```markdown
    # Implementation Plan: {feature-name}
 
@@ -204,3 +201,16 @@ If status is already `IMPLEMENTING`:
 - If coder fails on a task: Mark task as failed, ask user to continue or fix manually
 - If worktree creation fails: Continue without worktree, warn user
 - If spec not found: "Spec not found. Create with: /ab:spec {name}"
+
+## Note on Team Collaboration
+
+Since specs and plans are stored in `.auto-build/` (committed to git), any team member can:
+1. Clone the repo and see existing specs/plans
+2. Continue implementation from where another developer left off
+3. Share memory (patterns, gotchas) with the team
+
+After making progress, commit changes to share with team:
+```bash
+git add .auto-build/
+git commit -m "build: {feature-name} - completed task X/Y"
+```

plugin/commands/memory-save.md

@@ -101,7 +101,7 @@ For **codebase**:
 ### 5. Create Session Record
 
 Save comprehensive session record:
-`.auto-build-data/memory/sessions/{timestamp}-{feature}.json`
+`.auto-build/memory/sessions/{timestamp}-{feature}.json`
 
 ```json
 {
@@ -123,8 +123,8 @@ Save comprehensive session record:
 After saving to JSON files, convert and sync to CLAUDE.md-compatible format:
 
 1. **Load memory files**:
-   - Read `.auto-build-data/memory/patterns.json`
+   - Read `.auto-build/memory/patterns.json`
-   - Read `.auto-build-data/memory/gotchas.json`
+   - Read `.auto-build/memory/gotchas.json`
 
 2. **Convert to Markdown**:
    ```markdown

plugin/commands/memory-search.md

@@ -14,9 +14,9 @@ Search through accumulated patterns, gotchas, and session insights to find relev
 ## Search Scope
 
 Searches across:
-1. `.auto-build-data/memory/patterns.json`
+1. `.auto-build/memory/patterns.json`
-2. `.auto-build-data/memory/gotchas.json`
+2. `.auto-build/memory/gotchas.json`
-3. `.auto-build-data/memory/sessions/*.json`
+3. `.auto-build/memory/sessions/*.json`
 
 ## Workflow
 

plugin/commands/migrate.md

@@ -0,0 +1,141 @@
+---
+description: Migrate from old .auto-build-data/ structure to new .auto-build/ structure
+---
+
+# Migrate Auto-Build Data
+
+Migrate specs and memory from the old `.auto-build-data/` location to the new `.auto-build/` location for team collaboration.
+
+## When to Use
+
+Run this command if you have an existing project that was using Auto-Build v1.0.5 or earlier, where specs and memory were stored in `.auto-build-data/` (gitignored).
+
+## What It Does
+
+1. **Moves specs**: `.auto-build-data/specs/` → `.auto-build/specs/`
+2. **Moves memory**: `.auto-build-data/memory/` → `.auto-build/memory/`
+3. **Keeps local data**: `.auto-build-data/worktrees/` and `.auto-build-data/cache/` stay local
+4. **Updates .gitignore**: Ensures only `.auto-build-data/` is ignored
+
+## Workflow
+
+### 1. Check for Existing Data
+
+First, check if migration is needed:
+
+```bash
+# Check if old structure exists
+ls -la .auto-build-data/specs/
+ls -la .auto-build-data/memory/
+```
+
+If `.auto-build-data/specs/` or `.auto-build-data/memory/` exist with data, proceed with migration.
+
+### 2. Create New Structure
+
+```bash
+mkdir -p .auto-build/specs
+mkdir -p .auto-build/memory/sessions
+```
+
+### 3. Move Specs
+
+```bash
+# Move all feature specs
+if [ -d ".auto-build-data/specs" ] && [ "$(ls -A .auto-build-data/specs)" ]; then
+    mv .auto-build-data/specs/* .auto-build/specs/
+    echo "Moved specs to .auto-build/specs/"
+fi
+```
+
+### 4. Move Memory
+
+```bash
+# Move memory files
+if [ -f ".auto-build-data/memory/patterns.json" ]; then
+    mv .auto-build-data/memory/patterns.json .auto-build/memory/
+fi
+
+if [ -f ".auto-build-data/memory/gotchas.json" ]; then
+    mv .auto-build-data/memory/gotchas.json .auto-build/memory/
+fi
+
+if [ -d ".auto-build-data/memory/sessions" ] && [ "$(ls -A .auto-build-data/memory/sessions)" ]; then
+    mv .auto-build-data/memory/sessions/* .auto-build/memory/sessions/
+fi
+
+echo "Moved memory to .auto-build/memory/"
+```
+
+### 5. Initialize Missing Files
+
+If memory files don't exist, create them:
+
+```bash
+if [ ! -f ".auto-build/memory/patterns.json" ]; then
+    echo '{"patterns": [], "updated": null}' > .auto-build/memory/patterns.json
+fi
+
+if [ ! -f ".auto-build/memory/gotchas.json" ]; then
+    echo '{"gotchas": [], "updated": null}' > .auto-build/memory/gotchas.json
+fi
+```
+
+### 6. Update .gitignore
+
+Ensure `.gitignore` only ignores `.auto-build-data/`:
+
+```bash
+# Remove old entries if present
+grep -v "^\.auto-build-data/" .gitignore > .gitignore.tmp && mv .gitignore.tmp .gitignore
+
+# Add correct entry
+if ! grep -q "^\.auto-build-data/" .gitignore; then
+    echo "" >> .gitignore
+    echo "# Auto-Build local data (worktrees, cache)" >> .gitignore
+    echo ".auto-build-data/" >> .gitignore
+fi
+```
+
+### 7. Display Summary
+
+```
+════════════════════════════════════════════════════════════════
+  Migration Complete
+════════════════════════════════════════════════════════════════
+
+✅ Specs moved to .auto-build/specs/
+✅ Memory moved to .auto-build/memory/
+✅ .gitignore updated
+
+New structure (committed to git):
+  .auto-build/
+  ├── specs/{feature-name}/
+  │   ├── spec.md
+  │   ├── plan.md
+  │   └── status.json
+  └── memory/
+      ├── patterns.json
+      ├── gotchas.json
+      └── sessions/
+
+Local data (gitignored):
+  .auto-build-data/
+  ├── worktrees/
+  └── cache/
+
+NEXT STEPS:
+1. Review the migrated files
+2. Commit the new .auto-build/ directory:
+   git add .auto-build/
+   git commit -m "chore: Migrate auto-build specs and memory for team sharing"
+3. Push to share with your team
+
+════════════════════════════════════════════════════════════════
+```
+
+## Notes
+
+- This migration is **one-way** - the old structure is not preserved
+- After migration, specs and memory are **version controlled** and shared with your team
+- Worktrees remain local as they are machine-specific

plugin/commands/qa-review.md

@@ -21,7 +21,7 @@ Run iterative QA review with automatic fix attempts. The system will review code
 ### 1. Initialize
 
 - Determine feature from argument or find active build in status files
-- Load spec and plan from `.auto-build-data/specs/{feature-name}/`
+- Load spec and plan from `.auto-build/specs/{feature-name}/`
 - Create QA iteration directory: `.auto-build-data/cache/qa-iterations/{feature-name}/`
 - Set iteration counter to 0
 - Update status to `QA_REVIEW`
@@ -44,7 +44,7 @@ while iteration < MAX_ITERATIONS:
     - warning: Should fix
     - info: Nice to fix
 
-    Save to: .auto-build-data/cache/qa-iterations/{feature-name}/iteration-{n}.json
+    Save to: `.auto-build-data/cache/qa-iterations/{feature-name}/iteration-{n}.json`
 
     # 2c. Check Exit Condition
     if no errors and no warnings:
@@ -111,7 +111,7 @@ Next steps:
 
 ## Iteration Record Format
 
-`.auto-build-data/cache/qa-iterations/{feature-name}/iteration-{n}.json`:
+`.auto-build-data/cache/qa-iterations/{feature-name}/iteration-{n}.json` (local cache):
 ```json
 {
   "iteration": 1,

plugin/commands/spec.md

@@ -17,13 +17,13 @@ Create a comprehensive specification for a new feature using the spec-writer age
 ### 1. Initialize
 
 - Parse feature name from arguments (kebab-case: "User Dashboard" -> "user-dashboard")
-- Check if `.auto-build-data/` exists in the project root
+- Check if `.auto-build/` exists in the project root
   - If not: Create the directory structure:
     ```
-    mkdir -p .auto-build-data/specs
+    mkdir -p .auto-build/specs
-    mkdir -p .auto-build-data/memory
+    mkdir -p .auto-build/memory
     ```
-- Check if spec already exists at `.auto-build-data/specs/{feature-name}/`
+- Check if spec already exists at `.auto-build/specs/{feature-name}/`
   - If exists: Ask "Update existing spec or create new version?"
 
 ### 2. Gather Requirements
@@ -56,14 +56,14 @@ Launch the **spec-writer** agent with:
 ### 4. Generate Specification
 
 The spec-writer agent creates:
-- `.auto-build-data/specs/{feature-name}/spec.md` using the template
+- `.auto-build/specs/{feature-name}/spec.md` using the template
-- `.auto-build-data/specs/{feature-name}/status.json` with state: "SPEC_COMPLETE"
+- `.auto-build/specs/{feature-name}/status.json` with state: "SPEC_COMPLETE"
 
 ### 5. Confirmation
 
 Display:
 ```
-Specification created: .auto-build-data/specs/{feature-name}/spec.md
+Specification created: .auto-build/specs/{feature-name}/spec.md
 
 Summary:
 - [Key requirement 1]
@@ -74,8 +74,9 @@ Affected files identified: X files
 Estimated complexity: [Low/Medium/High]
 
 Next steps:
-- Review the spec: Read .auto-build-data/specs/{feature-name}/spec.md
+- Review the spec: Read .auto-build/specs/{feature-name}/spec.md
 - Start implementation: /ab:build {feature-name}
+- Commit to share with team: git add .auto-build/ && git commit -m "spec: {feature-name}"
 ```
 
 ## Spec Template

plugin/commands/status.md

@@ -9,11 +9,11 @@ Display the current status of Auto-Build, including active specs and builds.
 ## Workflow
 
 1. **Check Data Directory**
-   - Verify `.auto-build-data/` exists
+   - Verify `.auto-build/` exists
-   - If not, create it: `mkdir -p .auto-build-data/specs .auto-build-data/memory`
+   - If not, create it: `mkdir -p .auto-build/specs .auto-build/memory`
 
 2. **List Active Specs**
-   - Read all directories in `.auto-build-data/specs/`
+   - Read all directories in `.auto-build/specs/`
    - For each spec, read `status.json` to get current state
    - Display in table format:
      ```
@@ -28,7 +28,7 @@ Display the current status of Auto-Build, including active specs and builds.
    - Display active worktrees with their paths
 
 4. **Memory Summary**
-   - Count entries in `patterns.json` and `gotchas.json`
+   - Count entries in `.auto-build/memory/patterns.json` and `.auto-build/memory/gotchas.json`
    - Show last update timestamp
 
 ## Status States

plugin/hooks/post-install.sh

@@ -26,30 +26,43 @@ Last updated: Never
 
 (This file is automatically updated by /ab:memory-save)" > "$PROJECT_ROOT/.claude/rules/auto-build-memory.md"
 
-# 4. Create .auto-build-data/ structure
+# 4. Create .auto-build/ structure (COMMITTED TO GIT - shared with team)
-echo "Creating .auto-build-data/ structure..."
+echo "Creating .auto-build/ structure (shared with team)..."
-mkdir -p "$PROJECT_ROOT/.auto-build-data"/{specs,worktrees,memory/sessions,cache/qa-iterations}
+mkdir -p "$PROJECT_ROOT/.auto-build"/{specs,memory/sessions}
 
-# Initialize memory files
+# Initialize memory files if they don't exist
-echo '{"patterns": [], "updated": null}' > "$PROJECT_ROOT/.auto-build-data/memory/patterns.json"
+if [ ! -f "$PROJECT_ROOT/.auto-build/memory/patterns.json" ]; then
-echo '{"gotchas": [], "updated": null}' > "$PROJECT_ROOT/.auto-build-data/memory/gotchas.json"
+    echo '{"patterns": [], "updated": null}' > "$PROJECT_ROOT/.auto-build/memory/patterns.json"
+fi
+
+if [ ! -f "$PROJECT_ROOT/.auto-build/memory/gotchas.json" ]; then
+    echo '{"gotchas": [], "updated": null}' > "$PROJECT_ROOT/.auto-build/memory/gotchas.json"
+fi
+
+# 5. Create .auto-build-data/ structure (GITIGNORED - local only)
+echo "Creating .auto-build-data/ structure (local only)..."
+mkdir -p "$PROJECT_ROOT/.auto-build-data"/{worktrees,cache/qa-iterations}
+
+# Initialize worktree registry
+if [ ! -f "$PROJECT_ROOT/.auto-build-data/worktrees/worktree-registry.json" ]; then
     echo '{"worktrees": []}' > "$PROJECT_ROOT/.auto-build-data/worktrees/worktree-registry.json"
+fi
 
-# 5. Add .auto-build-data/ to .gitignore if not present
+# 6. Add .auto-build-data/ to .gitignore if not present
 if [ -f "$PROJECT_ROOT/.gitignore" ]; then
     if ! grep -q "^\.auto-build-data/" "$PROJECT_ROOT/.gitignore" 2>/dev/null; then
         echo "" >> "$PROJECT_ROOT/.gitignore"
-        echo "# Auto-Build runtime data (local only)" >> "$PROJECT_ROOT/.gitignore"
+        echo "# Auto-Build local data (worktrees, cache)" >> "$PROJECT_ROOT/.gitignore"
         echo ".auto-build-data/" >> "$PROJECT_ROOT/.gitignore"
         echo "Added .auto-build-data/ to .gitignore"
     fi
 else
-    echo "# Auto-Build runtime data (local only)" > "$PROJECT_ROOT/.gitignore"
+    echo "# Auto-Build local data (worktrees, cache)" > "$PROJECT_ROOT/.gitignore"
     echo ".auto-build-data/" >> "$PROJECT_ROOT/.gitignore"
     echo "Created .gitignore with .auto-build-data/ entry"
 fi
 
-# 6. Initialize or update CLAUDE.md
+# 7. Initialize or update CLAUDE.md
 if [ ! -f "$PROJECT_ROOT/CLAUDE.md" ]; then
     echo "Generating CLAUDE.md from template..."
     cp "$PLUGIN_DIR/templates/CLAUDE.md.template" "$PROJECT_ROOT/CLAUDE.md"
@@ -75,7 +88,8 @@ echo ""
 echo "What was installed:"
 echo "  ✅ .claude/rules/auto-build-patterns.md"
 echo "  ✅ .claude/rules/auto-build-memory.md"
-echo "  ✅ .auto-build-data/ structure"
+echo "  ✅ .auto-build/ (specs & memory - shared with team)"
+echo "  ✅ .auto-build-data/ (local worktrees & cache)"
 echo "  ✅ CLAUDE.md (created or updated)"
 echo "  ✅ .gitignore updated"
 echo ""
@@ -84,3 +98,5 @@ echo "  1. Restart Claude Code to load CLAUDE.md"
 echo "  2. Run /ab:help for available commands"
 echo "  3. Create your first spec: /ab:spec \"My Feature\""
 echo ""
+echo "Note: Don't forget to commit .auto-build/ to share specs with your team!"
+echo ""

plugin/rules/auto-build-patterns.md

@@ -10,8 +10,9 @@ These rules apply to all Auto-Build related files and workflows.
 
 ### 1. Portability First
 - All paths must be relative or dynamically determined
-- No hardcoded project-specific values in .auto-build/
+- No hardcoded project-specific values
-- Data lives in .auto-build-data/ which is gitignored
+- Shared data (specs, plans, memory) lives in `.auto-build/` (git-tracked)
+- Local data (worktrees, cache) lives in `.auto-build-data/` (gitignored)
 
 ### 2. Minimal Footprint
 - Never modify files outside the current task scope
@@ -29,7 +30,7 @@ These rules apply to all Auto-Build related files and workflows.
 - Always analyze the codebase before writing spec
 - Include at least 3 acceptance criteria
 - Identify affected files explicitly
-- Save to .auto-build-data/specs/{name}/
+- Save to `.auto-build/specs/{name}/` (git-tracked, shared with team)
 
 ### When Running /ab:build
 - Check for existing spec before proceeding
@@ -89,17 +90,21 @@ SPEC_DRAFT → SPEC_COMPLETE → PLANNING → PLANNING_COMPLETE →
 IMPLEMENTING → IMPLEMENTATION_COMPLETE → QA_REVIEW → COMPLETE
 ```
 
-### Memory Files
+### Memory Files (`.auto-build/memory/` - git-tracked)
-- patterns.json: Reusable code patterns
+- patterns.json: Reusable code patterns (shared with team)
-- gotchas.json: Problems and solutions
+- gotchas.json: Problems and solutions (shared with team)
-- codebase-map.json: Structural insights
+- sessions/*.json: Session history (shared with team)
-- sessions/*.json: Session history
+
+### Local Data Files (`.auto-build-data/` - gitignored)
+- worktrees/worktree-registry.json: Local worktree tracking
+- cache/qa-iterations/*.json: Temporary QA cache
 
 ## Git Worktree Rules
 
 - Create worktrees outside project: ../ab-worktrees/
 - Branch naming: feature/ab-{feature-name}
-- Registry tracking in worktree-registry.json
+- Registry tracking in `.auto-build-data/worktrees/worktree-registry.json`
+- No symlinks needed - specs/plans/memory are in git, available in all worktrees
 - Cleanup only merged branches automatically
 
 ## Error Handling

plugin/scripts/worktree-create.sh

@@ -66,18 +66,8 @@ fi
 echo "Creating worktree..."
 git worktree add "$WORKTREE_PATH" "$BRANCH_NAME"
 
-# Create symlink to .auto-build-data/ from main repo
+# Note: No symlinks needed - specs and memory are in .auto-build/ (git-tracked)
-# This ensures specs, plans, and status are shared across worktrees
+# The worktree automatically has access to all shared data
-if [ -d "$PROJECT_ROOT/.auto-build-data" ]; then
-    echo "Creating symlink to .auto-build-data/..."
-    ln -s "$PROJECT_ROOT/.auto-build-data" "$WORKTREE_PATH/.auto-build-data"
-fi
-
-# Also symlink .claude/rules/ for memory access
-if [ -d "$PROJECT_ROOT/.claude/rules" ]; then
-    mkdir -p "$WORKTREE_PATH/.claude"
-    ln -s "$PROJECT_ROOT/.claude/rules" "$WORKTREE_PATH/.claude/rules"
-fi
 
 # Update registry
 REGISTRY_FILE="$PROJECT_ROOT/.auto-build-data/worktrees/worktree-registry.json"