Production-ready Claude Code configuration for AI/ML engineers with 85 agents, 37 skills, 65 rules, and enterprise-grade security hooks.
- 85 Specialized Agents - AI/ML, infrastructure, security, frameworks
- 37 Skills - Reusable workflows and automation
- 65 Language Rules - TypeScript, Python, Go, Rust, and more
- 5 Security Hooks - Trail of Bits best practices
- Multi-Terminal Parallel Execution - Speed up development 3x
- Windows Compatible - All hooks work without jq dependency
# Clone the repository
git clone https://github.com/Arnav-0/claude-code-setup.git
# Run the install script
cd claude-code-setup
./install.sh
# Or manual install
cp -r agents ~/.claude/
cp -r skills ~/.claude/
cp -r rules ~/.claude/
cp -r commands ~/.claude/
cp -r hooks ~/.claude/
cp -r agent-teams ~/.claude/
cp settings.json ~/.claude/
cp CLAUDE.md ~/.claude/
cp SYSTEM-GUIDE.md ~/.claude/
cp USAGE-GUIDE.md ~/.claude/
# Restart Claude Code
claudeclaude-code-setup/
├── agents/ # 85 AI agents
│ ├── idea-researcher.md # Deep research & ideas
│ ├── project-planner.md # Multi-terminal planning
│ ├── elite-builder.md # FAANG-quality code
│ ├── ml-architect.md # ML system design
│ ├── fullstack-ml.md # AI app builder
│ ├── innovation-tracker.md # Track AI news
│ └── ... (79 more)
├── skills/ # 37 reusable skills
├── rules/ # 65 language rules
│ ├── common/ # Universal principles
│ ├── typescript/ # TypeScript specific
│ ├── python/ # Python specific
│ └── ...
├── commands/ # 10 slash commands
│ ├── pr.md # Create pull requests
│ ├── tdd.md # Test-driven development
│ ├── review.md # Multi-perspective review
│ └── ...
├── hooks/ # 5 security hooks
│ ├── block-dangerous.sh
│ ├── protect-credentials.sh
│ ├── audit-log.sh
│ ├── anti-rationalization.sh
│ └── session-context.sh
├── agent-teams/ # Pre-configured teams
├── settings.json # Full configuration
├── CLAUDE.md # Quick reference
├── SYSTEM-GUIDE.md # Complete guide
└── USAGE-GUIDE.md # AI/ML workflows
| Agent | Purpose |
|---|---|
idea-researcher |
Deep research, market analysis, idea validation |
project-planner |
Creates multi-terminal parallel execution plans |
elite-builder |
FAANG-quality, production-grade implementation |
ml-architect |
End-to-end ML system design |
fullstack-ml |
Complete AI applications (FastAPI + React + ML) |
innovation-tracker |
Track Claude features & AI industry news |
Use the idea-researcher agent to explore opportunities in AI agents
Use the project-planner agent to plan my RAG application
Use the elite-builder agent to implement the authentication system
Use the ml-architect agent to design the ML pipeline
| Command | Purpose |
|---|---|
/pr |
Create pull request with analysis |
/tdd |
Test-driven development workflow |
/review |
Multi-perspective code review |
/ultrathink |
Deep analysis for complex problems |
/debug |
Systematic debugging workflow |
/refactor |
Safe refactoring with tests |
/quality-build |
Full QA pipeline |
| Protection | Action |
|---|---|
rm -rf commands |
Blocked, suggests safer alternatives |
| Push to main/master | Blocked, requires feature branch |
| Force push | Blocked, suggests --force-with-lease |
| .env file writes | Blocked, use environment variables |
| Credential access | Blocked (SSH, AWS, GPG keys) |
| Premature completion | Blocked by anti-rationalization |
~/.ssh/* # SSH keys
~/.aws/credentials # AWS secrets
~/.gnupg/* # GPG keys
~/.kube/config # Kubernetes config
**/.env # Environment files
**/secrets.json # Secret files
All mutations logged to ~/.claude/logs/audit-YYYY-MM-DD.jsonl
git worktree add ../project-backend -b feature/backend
git worktree add ../project-frontend -b feature/frontend
git worktree add ../project-ml -b feature/ml-pipeline# Terminal 1
cd ../project-backend && claude -p "Implement backend API..."
# Terminal 2
cd ../project-frontend && claude -p "Implement frontend..."
# Terminal 3
cd ../project-ml && claude -p "Implement ML pipeline..."git checkout main
git merge feature/backend feature/frontend feature/ml-pipeline| Mode | Trigger | Use Case |
|---|---|---|
| Basic | think |
Simple analysis |
| Medium | think hard |
Complex problems |
| Maximum | ultrathink |
Architecture decisions |
| Plan | Shift+Tab |
Before implementation |
architect, code-reviewer, debugger, planner, tdd-guide,
build-error-resolver, refactor-cleaner, doc-updater, e2e-runner, researcher
ai-engineer, ml-engineer, data-scientist, data-engineer,
nlp-engineer, computer-vision-expert, reinforcement-learning-expert, mlops-engineer
typescript-reviewer, python-reviewer, go-reviewer, rust-reviewer,
java-reviewer, kotlin-reviewer, cpp-reviewer, flutter-reviewer
devops-engineer, cloud-architect, kubernetes-expert, terraform-expert,
ansible-expert, aws-infrastructure-expert, azure-infrastructure-expert,
gcp-infrastructure-expert, cicd-pipeline-expert, gitops-expert,
monitoring-expert, observability-expert
react-expert, nextjs-expert, vue-expert, angular-expert,
svelte-expert, django-expert, fastapi-expert, rails-expert,
spring-expert, nestjs-expert, graphql-expert, grpc-expert, websocket-expert
security-reviewer, security-penetration-tester, devsecops-engineer, zero-trust-architect
database-architect, database-reviewer, postgresql-expert, mongodb-expert,
elasticsearch-expert, redis-expert, neo4j-expert
1. "Use the idea-researcher agent to explore opportunities in [domain]"
2. "Use the project-planner agent to create a multi-terminal plan"
3. Open 3 terminals and execute workstreams in parallel
4. "Use the elite-builder agent to implement [critical feature]"
5. "Use the code-reviewer agent to review all changes"
6. /pr
1. "Use the debugger agent to investigate [error]"
2. "think hard about the root cause"
3. "Use the tdd-guide agent to write a regression test"
4. Fix the bug
5. /review
1. "Use the ml-architect agent to design a system for [problem]"
2. Review the architecture
3. "Use the project-planner agent to break it into tasks"
4. "Use the fullstack-ml agent to implement the MVP"
For enhanced audit logging (not required):
# Windows
winget install jqlang.jq
# macOS
brew install jq
# Linux
sudo apt install jq- Trail of Bits claude-code-config - Security hooks
- rohitg00/awesome-claude-code-toolkit - Agents & skills
- affaan-m/everything-claude-code - Configuration patterns
- disler/claude-code-hooks-mastery - Hook patterns
MIT License - Feel free to use, modify, and distribute.
Built for maximum potential. Ship amazing AI products!