Enhanced code review and frontend validation with intelligent triggers: Code Review Agent Enhancement: - Added Sequential Thinking MCP integration for complex issues - Triggers on 2+ rejections or 3+ critical issues - New escalation format with root cause analysis - Comprehensive solution strategies with trade-off evaluation - Educational feedback to break rejection cycles - Files: .claude/agents/code-review.md (+308 lines) - Docs: CODE_REVIEW_ST_ENHANCEMENT.md, CODE_REVIEW_ST_TESTING.md Frontend Design Skill Enhancement: - Automatic invocation for ANY UI change - Comprehensive validation checklist (200+ checkpoints) - 8 validation categories (visual, interactive, responsive, a11y, etc.) - 3 validation levels (quick, standard, comprehensive) - Integration with code review workflow - Files: .claude/skills/frontend-design/SKILL.md (+120 lines) - Docs: UI_VALIDATION_CHECKLIST.md (462 lines), AUTOMATIC_VALIDATION_ENHANCEMENT.md (587 lines) Settings Optimization: - Repaired .claude/settings.local.json (fixed m365 pattern) - Reduced permissions from 49 to 33 (33% reduction) - Removed duplicates, sorted alphabetically - Created SETTINGS_PERMISSIONS.md documentation Checkpoint Command Enhancement: - Dual checkpoint system (git + database) - Saves session context to API for cross-machine recall - Includes git metadata in database context - Files: .claude/commands/checkpoint.md (+139 lines) Decision Rationale: - Sequential Thinking MCP breaks rejection cycles by identifying root causes - Automatic frontend validation catches UI issues before code review - Dual checkpoints enable complete project memory across machines - Settings optimization improves maintainability Total: 1,200+ lines of documentation and enhancements Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
16 KiB
AutoCoder Resources Integration Guide
Date: 2026-01-17 Source: AutoCoder project (Autocode-remix fork) Status: Successfully integrated
Overview
This guide documents the AutoCoder resources that have been integrated into ClaudeTools, including commands, skills, templates, and an MCP server for feature management.
What was extracted:
- 2 Commands (create-spec, checkpoint)
- 1 Skill (frontend-design)
- 4 Templates (app spec, coding prompts, initializer)
- 1 MCP Server (feature management)
Purpose: Enable autonomous coding workflows with spec-driven development and feature tracking.
Directory Structure
D:\ClaudeTools/
├── .claude/
│ ├── commands/
│ │ ├── sync.md # EXISTING - Cross-machine sync
│ │ ├── create-spec.md # NEW - Create app specification
│ │ └── checkpoint.md # NEW - Create development checkpoint
│ │
│ ├── skills/
│ │ └── frontend-design/ # NEW - Frontend design skill
│ │ ├── SKILL.md
│ │ └── LICENSE.txt
│ │
│ └── templates/ # NEW directory
│ ├── app_spec.template.txt
│ ├── coding_prompt.template.md
│ ├── coding_prompt_yolo.template.md
│ └── initializer_prompt.template.md
│
└── mcp-servers/ # NEW directory
└── feature-management/
├── feature_mcp.py # MCP server implementation
├── __init__.py
├── README.md # Documentation
└── config.example.json # Configuration example
1. Commands
create-spec.md
Purpose: Create a comprehensive application specification for autonomous coding
How to use:
# In Claude Code
/create-spec
# Claude will guide you through creating:
# - Project overview
# - Tech stack
# - Features list
# - Database schema
# - API endpoints
# - UI/UX requirements
Output: Creates APP_SPEC.md in project root with full specification
When to use:
- Starting a new autonomous coding project
- Documenting requirements for an agent-driven build
- Creating structured input for the feature management system
File location: D:\ClaudeTools\.claude\commands\create-spec.md
checkpoint.md
Purpose: Create a detailed development checkpoint with commit and summary
How to use:
# In Claude Code
/checkpoint
# Claude will:
# 1. Analyze recent changes
# 2. Create a detailed commit message
# 3. Commit changes with co-authorship
# 4. Save session context to context recall system
Output:
- Git commit with detailed message
- Context saved to conversation_contexts table
- Decision logged in decision_logs table
When to use:
- After completing a feature
- Before switching to a different task
- At natural breakpoints in development
File location: D:\ClaudeTools\.claude\commands\checkpoint.md
2. Skills
frontend-design
Purpose: Specialized skill for creating modern, production-ready frontend designs
How to use:
# In Claude Code
/frontend-design
# Claude will use the skill to:
# - Design responsive layouts
# - Create component hierarchies
# - Implement modern UI patterns
# - Follow accessibility best practices
Features:
- Modern framework patterns (React, Vue, Svelte)
- Responsive design (mobile-first)
- Accessibility (ARIA, semantic HTML)
- Performance optimization
- Component reusability
File location: D:\ClaudeTools\.claude\skills\frontend-design\
3. Templates
app_spec.template.txt
Purpose: Template for creating application specifications
Usage: Used by /create-spec command to structure the output
Contents:
- Project metadata
- Technology stack
- Feature categories
- Database schema
- API endpoints
- Authentication/authorization
- Deployment requirements
File location: D:\ClaudeTools\.claude\templates\app_spec.template.txt
coding_prompt.template.md
Purpose: Standard coding prompt for autonomous agents
Usage: Structured prompt that defines:
- Agent role and capabilities
- Development workflow
- Quality standards
- Testing requirements
- Error handling patterns
When to use: Starting an autonomous coding agent session
File location: D:\ClaudeTools\.claude\templates\coding_prompt.template.md
coding_prompt_yolo.template.md
Purpose: Aggressive "move fast" coding prompt for rapid prototyping
Usage: Alternative prompt that prioritizes:
- Speed over perfect code
- Getting features working quickly
- Minimal testing (just basics)
- Iterative refinement
When to use:
- Proof of concepts
- Rapid prototyping
- MVP development
- Hackathons
File location: D:\ClaudeTools\.claude\templates\coding_prompt_yolo.template.md
initializer_prompt.template.md
Purpose: Prompt for initializing a new project from specification
Usage: Sets up:
- Project structure
- Dependencies
- Configuration files
- Initial feature list
- Database setup
When to use: Starting a new project from an APP_SPEC.md
File location: D:\ClaudeTools\.claude\templates\initializer_prompt.template.md
4. MCP Server - Feature Management
Overview
The Feature Management MCP Server provides native Claude Code integration for managing autonomous coding workflows with a priority-based feature queue.
Setup
Step 1: Install Dependencies
# Activate ClaudeTools virtual environment
D:\ClaudeTools\venv\Scripts\activate
# Install required packages
pip install fastmcp sqlalchemy pydantic
Step 2: Configure Claude Desktop
Edit Claude Desktop config file:
- Windows:
%APPDATA%\Claude\claude_desktop_config.json - macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Linux:
~/.config/claude/claude_desktop_config.json
Add this configuration:
{
"mcpServers": {
"features": {
"command": "python",
"args": ["D:\\ClaudeTools\\mcp-servers\\feature-management\\feature_mcp.py"],
"env": {
"PROJECT_DIR": "D:\\ClaudeTools\\projects\\your-project"
}
}
}
}
Step 3: Restart Claude Desktop
Close and reopen Claude Desktop for changes to take effect.
Step 4: Verify Installation
In Claude Code, you should now have access to these MCP tools:
feature_get_statsfeature_get_nextfeature_mark_passingfeature_mark_in_progressfeature_skipfeature_clear_in_progressfeature_get_for_regressionfeature_create_bulk
MCP Server Tools Reference
Quick Reference:
| Tool | Purpose | Usage |
|---|---|---|
feature_get_stats |
Progress overview | Start/end of session |
feature_get_next |
Get next feature | Start implementation |
feature_mark_in_progress |
Claim feature | After getting next |
feature_mark_passing |
Complete feature | After implementation |
feature_skip |
Defer feature | When blocked |
feature_clear_in_progress |
Reset status | When abandoning |
feature_get_for_regression |
Get test features | After changes |
feature_create_bulk |
Initialize features | Project setup |
Full documentation: See D:\ClaudeTools\mcp-servers\feature-management\README.md
Typical Autonomous Coding Workflow
Phase 1: Project Initialization
# 1. Create app specification
/create-spec
# Follow prompts to define your application
# 2. Review and save APP_SPEC.md
# Edit as needed, commit to version control
# 3. Initialize feature list
# Use MCP tool to create features from spec
feature_create_bulk(features=[...])
Phase 2: Development Loop
# 1. Check progress
feature_get_stats()
# Output: 15/50 features (30%)
# 2. Get next feature
next_feature = feature_get_next()
# Feature #16: "User authentication endpoint"
# 3. Mark in-progress
feature_mark_in_progress(feature_id=16)
# 4. Implement feature
# ... write code, run tests ...
# 5. Create checkpoint
/checkpoint
# Commits changes and saves context
# 6. Mark complete
feature_mark_passing(feature_id=16)
# 7. Regression test
regression = feature_get_for_regression(limit=5)
# Test 5 random passing features
Phase 3: Handling Blockers
# Get next feature
next = feature_get_next()
# Feature #20: "OAuth integration"
# Realize it depends on incomplete feature #25
feature_skip(feature_id=20)
# Moved to end of queue
# Continue with next available
next = feature_get_next()
# Feature #21: "Email validation"
Integration with ClaudeTools API
The AutoCoder resources integrate seamlessly with the existing ClaudeTools infrastructure:
1. Context Recall Integration
Save feature completion context:
POST /api/conversation-contexts
{
"project_id": "uuid",
"context_type": "feature_completion",
"title": "Completed Feature #16: User authentication",
"dense_summary": "Implemented JWT-based auth with bcrypt password hashing...",
"relevance_score": 8.0,
"tags": ["authentication", "feature-16", "jwt", "bcrypt"]
}
2. Decision Logging
Log architectural decisions:
POST /api/decision-logs
{
"project_id": "uuid",
"decision_type": "technical",
"decision_text": "Use JWT for stateless authentication",
"rationale": "Scalability, no server-side session storage needed",
"alternatives_considered": ["Session cookies", "OAuth only"],
"impact": "high",
"tags": ["authentication", "architecture"]
}
3. Session Tracking
Track feature development sessions:
POST /api/sessions
{
"project_id": "uuid",
"machine_id": "machine-uuid",
"started_at": "2026-01-17T10:00:00Z",
"metadata": {"feature_id": 16, "feature_name": "User authentication"}
}
4. Problem Solutions
Save implementation solutions:
POST /api/context-snippets
{
"snippet_type": "solution",
"title": "JWT token validation middleware",
"content": "async def validate_token(request): ...",
"language": "python",
"tags": ["authentication", "jwt", "middleware"],
"relevance_score": 7.5
}
Configuration Files
Claude Desktop Config
Full example configuration:
{
"mcpServers": {
"features": {
"command": "D:\\ClaudeTools\\venv\\Scripts\\python.exe",
"args": ["D:\\ClaudeTools\\mcp-servers\\feature-management\\feature_mcp.py"],
"env": {
"PROJECT_DIR": "D:\\ClaudeTools\\projects\\my-web-app"
}
}
},
"globalShortcut": "Ctrl+Space"
}
Environment Variables
For MCP server:
PROJECT_DIR- Required. Where features.db will be stored.
For ClaudeTools API:
DATABASE_URL- Connection string to ClaudeTools databaseJWT_SECRET_KEY- For API authenticationENCRYPTION_KEY- For credential encryption
Testing the Integration
1. Test Commands
# Test create-spec
/create-spec
# Should display specification creation interface
# Test checkpoint
/checkpoint
# Should create git commit and save context
2. Test Skills
# Test frontend-design
/frontend-design
# Should activate frontend design mode
3. Test MCP Server
# In Claude Code with MCP server running
# Test stats
feature_get_stats()
# Should return progress statistics
# Test get next
feature_get_next()
# Should return next feature or error if empty
4. Test Templates
Templates are used internally by commands, but you can view them:
# View templates
cat D:\ClaudeTools\.claude\templates\app_spec.template.txt
cat D:\ClaudeTools\.claude\templates\coding_prompt.template.md
Troubleshooting
Commands Not Available
Problem: /create-spec or /checkpoint not showing in command list
Solution:
- Verify files exist in
.claude/commands/ - Restart Claude Code
- Check file permissions (should be readable)
Skill Not Loading
Problem: /frontend-design skill not available
Solution:
- Verify
SKILL.mdexists in.claude/skills/frontend-design/ - Check SKILL.md syntax (must be valid markdown)
- Restart Claude Code
MCP Server Not Connecting
Problem: Feature tools not available in Claude Code
Solution:
- Verify Claude Desktop config is valid JSON
- Check
PROJECT_DIRenvironment variable is set - Ensure Python can be found (use full path if needed)
- Check MCP server logs (see Claude Desktop logs)
- Restart Claude Desktop (not just the window)
Windows log location:
%APPDATA%\Claude\logs\
macOS log location:
~/Library/Logs/Claude/
Database Issues
Problem: MCP server can't create/access database
Solution:
- Verify
PROJECT_DIRexists and is writable - Check file permissions on
PROJECT_DIR - Manually create directory if needed:
mkdir -p "D:\ClaudeTools\projects\your-project"
Best Practices
1. Spec-Driven Development
Always start with a specification:
- Use
/create-specto document requirements - Review and refine the spec before coding
- Use spec as input for
feature_create_bulk - Keep spec updated as requirements evolve
2. Checkpoint Frequently
Create checkpoints at natural boundaries:
- After completing each feature
- Before starting risky refactoring
- At end of coding sessions
- When switching between tasks
3. Feature Management
Maintain clean feature state:
- Always mark features in-progress when starting
- Mark passing only when fully tested
- Skip features when blocked (don't leave them in-progress)
- Use regression testing after significant changes
4. Context Recall
Integrate with ClaudeTools context system:
- Save feature completions to conversation_contexts
- Log architectural decisions to decision_logs
- Store reusable solutions to context_snippets
- Tag everything for easy retrieval
Migration Notes
From AutoCoder to ClaudeTools
What changed:
- Commands moved from AutoCoder
.claude/to ClaudeTools.claude/ - MCP server moved to dedicated
mcp-servers/directory - Templates now in
.claude/templates/(new directory) - Skills now in
.claude/skills/(new directory)
What stayed the same:
- Command syntax and functionality
- MCP server API (same tools, same parameters)
- Template structure
- Skill format
Backwards Compatibility
These AutoCoder resources are compatible with:
- Claude Code (current version)
- Claude Desktop MCP protocol
- ClaudeTools API (Phase 6, context recall)
Not compatible with:
- Older AutoCoder versions (pre-MCP)
- Legacy JSON-based feature tracking (auto-migrated)
Next Steps
Recommended Actions
-
Try the commands:
- Run
/create-specon a test project - Create a checkpoint with
/checkpoint
- Run
-
Set up MCP server:
- Configure Claude Desktop
- Test feature management tools
- Create initial feature list
-
Integrate with ClaudeTools:
- Connect feature completions to context recall
- Log decisions to decision_logs
- Track sessions with metadata
-
Customize templates:
- Review templates in
.claude/templates/ - Adjust to match your coding style
- Add project-specific requirements
- Review templates in
Resources
Documentation
- MCP Server:
mcp-servers/feature-management/README.md - Config Example:
mcp-servers/feature-management/config.example.json - ClaudeTools Docs:
.claude/CLAUDE.md - Context Recall:
.claude/CONTEXT_RECALL_QUICK_START.md
Source Files
- Commands:
.claude/commands/ - Skills:
.claude/skills/ - Templates:
.claude/templates/ - MCP Servers:
mcp-servers/
AutoCoder Project
- Original Source:
/c/Users/MikeSwanson/claude-projects/Autocode-remix/Autocode-fork/autocoder-master - Conversation History:
imported-conversations/auto-claude-variants/autocode-remix-fork/
Version History
| Date | Version | Changes |
|---|---|---|
| 2026-01-17 | 1.0 | Initial integration from AutoCoder |
Last Updated: 2026-01-17 Integration Status: Complete Tested: Windows 11, ClaudeTools v0.95