Awesome Claude Code
A curated list of tools, plugins, MCP servers, and resources for Claude Code.
Know something that should be on this list? Submit a PR or open an issue.
Plugins
| Plugin | What It Does |
|---|---|
| OMC (Oh My Claude Code) | Advanced session orchestration: autopilot, parallel agents, persistence loops, smart model routing |
| BMAD | Multi-agent roles: Architect, Developer, QA, Security Auditor, Product Manager |
| Context7 | AI-powered documentation search — fetches up-to-date library docs and code examples |
| PR Review Toolkit | Comprehensive PR review with specialized analysis agents |
| Commit Commands | Streamlined git commit, push, and PR workflows |
Companion CLIs
CLI tools that wrap Claude Code (and sometimes Codex / Gemini) to extend its reach beyond a single session.
| Tool | What It Does |
|---|---|
| Tribune | Convenes three advocates — Proposer, Skeptic, Red Team — on a hard decision and commits the argument as an ADR. Runs off your Claude Code / Codex / Gemini CLI subscriptions, no API keys. Ships a /tribune slash command for Claude Code sessions. |
MCP Servers
See docs/mcp-servers.md for detailed setup and configuration.
| Server | Purpose | Link |
|---|---|---|
| Playwright | Browser automation, screenshots, UI testing | npm |
| Memory | Persistent knowledge graph across sessions | npm |
| Filesystem | Extended file operations | npm |
| PostgreSQL | Database queries and schema inspection | npm |
| SQLite | Local database operations | npm |
| Puppeteer | Web scraping and headless browser | npm |
| Brave Search | Web search from Claude Code sessions | npm |
| GitHub | Extended GitHub operations beyond gh CLI | npm |
Agent Skills Ecosystem
The Skills CLI (npx skills) is the package manager for the open agent skills ecosystem. Install community-maintained skills that teach Claude Code best practices for specific frameworks and tools.
npx skills find "react" # Search
npx skills add vercel-labs/agent-skills@react-best-practices -y # Install
npx skills update # Update all
See docs/skills-ecosystem.md for the full guide — evaluation criteria, recommended skills by stack, redundancy cleanup, and creating your own.
Top skills by category:
| Category | Skill | Installs |
|---|---|---|
| React | vercel-labs/agent-skills@react-best-practices | 241K |
| Azure | microsoft/azure-skills@azure-prepare | 81K |
| Tailwind | wshobson/agents@tailwind-design-system | 23K |
| Playwright | currents-dev/playwright-best-practices-skill | 14.8K |
| Playwright CLI | microsoft/playwright-cli | 10.3K |
| NestJS | kadajett/agent-nestjs-skills@nestjs-best-practices | 8.8K |
| Accessibility | addyosmani/web-quality-skills@accessibility | 4K |
| Performance | sickn33/antigravity-awesome-skills@web-performance-optimization | 1.3K |
| OWASP Security | hoodini/ai-agents-skills@owasp-security | 716 |
Community Resources
| Resource | Description |
|---|---|
| Claude Code Documentation | Official Anthropic docs for Claude Code |
| Claude Code GitHub | Official repo — issues, releases, and discussions |
| MCP Specification | The Model Context Protocol specification and guides |
| MCP Servers Directory | Official directory of MCP servers |
| Anthropic Cookbook | Recipes and patterns for building with Claude |
CLAUDE.md Templates
This playbook includes 9 templates. See templates/ for the full list:
- General / TypeScript, React / Next.js, Node.js API, Python, Full-stack monorepo, Go, Rust, React Native / Mobile, DevOps / Infrastructure
Tips & Tricks
/clearis your best friend. When switching tasks, always clear context. One task per session = better results.- Paste real errors. Don’t describe the error — paste the stack trace. Claude can parse it faster than you can explain it.
- Scope-lock your prompts. “Only modify
src/auth/” prevents Claude from wandering into unrelated code. - Use reverse prompting. “Ask me 20 questions before starting” produces better specs than writing them yourself.
- Model routing saves money. Haiku for searches, Sonnet for code, Opus for architecture. See model comparison.
- Hooks catch mistakes early. A 30-second TypeScript check after every edit saves 10-minute build failures later.
This list is maintained by the community. Contributions welcome.