<div align="center">

🤖 Awesome AI Coding Agent Tools

Tools, libraries, MCP servers, and frameworks that power AI coding agents.

---

✨ Featured Tools

<p> <a href="https://github.com/modelcontextprotocol/servers"><img src="https://img.shields.io/github/stars/modelcontextprotocol/servers?style=for-the-badge&logo=github&logoColor=white&label=MCP%20Servers&color=0080FF" alt="MCP Servers" /></a> <a href="https://github.com/anthropics/skills"><img src="https://img.shields.io/github/stars/anthropics/skills?style=for-the-badge&logo=anthropic&logoColor=white&label=Anthropic%20Skills&color=D4A574" alt="Anthropic Skills" /></a> <a href="https://github.com/langchain-ai/langchain"><img src="https://img.shields.io/github/stars/langchain-ai/langchain?style=for-the-badge&logo=langchain&logoColor=white&label=LangChain&color=1C3A3A" alt="LangChain" /></a> <a href="https://github.com/langchain-ai/langgraph"><img src="https://img.shields.io/github/stars/langchain-ai/langgraph?style=for-the-badge&logo=langchain&logoColor=white&label=LangGraph&color=00C896" alt="LangGraph" /></a> </p> <p> <a href="https://github.com/vercel/ai"><img src="https://img.shields.io/github/stars/vercel/ai?style=for-the-badge&logo=vercel&logoColor=white&label=Vercel%20AI&color=000000" alt="Vercel AI SDK" /></a> <a href="https://github.com/e2b-dev/E2B"><img src="https://img.shields.io/github/stars/e2b-dev/E2B?style=for-the-badge&logo=python&logoColor=white&label=E2B&color=FF6B35" alt="E2B" /></a> <a href="https://github.com/langfuse/langfuse"><img src="https://img.shields.io/github/stars/langfuse/langfuse?style=for-the-badge&logo=github&logoColor=white&label=Langfuse&color=6F42C1" alt="Langfuse" /></a> <a href="https://github.com/yamadashy/repomix"><img src="https://img.shields.io/github/stars/yamadashy/repomix?style=for-the-badge&logo=github&logoColor=white&label=Repomix&color=0066CC" alt="Repomix" /></a> </p> </div>

---

This list collects the tools, libraries, MCP servers, and frameworks that power AI coding agents — from the skills and plugins that extend them, to the SDKs you'd build them with, to the supporting layers (memory, sandboxing, evals, observability, benchmarks) that turn demos into production. The agents themselves (Claude Code, Cursor, aider, Cline, …) live in their own repos; this list is the ecosystem around them. Inclusion is based on visible community use rather than star counts. Every entry includes nested Strengths and Caveats bullets so you can compare at a glance.

Contents

• Extensions, Skills & Rules
  • Claude Code Skills & Plugins
  • Cursor Rules
  • MCP Servers
  • Curated Awesome Lists
  • Methodologies & Spec-Driven Workflows
  • Skill & Plugin Marketplaces
• Agent SDKs & Frameworks
  • Official Vendor SDKs
  • Orchestration Frameworks
  • Lightweight & Typed
  • Structured Output & Multi-Provider
  • Stateful Agents
• Supporting Infrastructure
  • Memory
  • Vector Databases
  • Codebase Context
  • Sandboxing & Execution
  • Workflow & Pipeline Orchestration
  • Evaluation
  • Observability
  • Prompt Management
  • Benchmarks & Leaderboards
  • Test Generation
  • Local LLM Runners
• Learning Resources
  • Articles & Blogs
  • Tutorials & Courses
  • Books
  • Podcasts
  • Datasets
• Community

Extensions, Skills & Rules

Add-ons that extend AI coding agents with new capabilities, knowledge, or rules.

Claude Code Skills & Plugins

• Anthropic Claude Cookbooks - Notebooks and recipes for Claude skills, patterns, tool use, and memory management.
  • Strengths: 95+ Python Jupyter examples; RAG/classification/summarization; multimodal vision.
  • Caveats: Python-centric; assumes Claude API familiarity; rapid API changes.
• Anthropic Claude Plugins Official - Anthropic's official directory of curated Claude Code plugins.
  • Strengths: curated official plugins; simple install via CLI; standardized plugin structure.
  • Caveats: trust plugins before installing; limited quality control for external plugins; verify independently.
• Anthropic Claude Quickstarts - Deployable starter projects to jumpstart Claude API applications.
  • Strengths: ready-to-use templates; covers support/finance/automation/coding; well-documented.
  • Caveats: requires Claude API key; per-project setup; no production guarantees.
• Anthropic Skills - Official open standard repo for Agent Skills in SKILL.md format.
  • Strengths: simple folder-based structure; multi-platform (Code/AI/API); 131 production examples.
  • Caveats: educational/demo purpose; behavior variability vs documentation; testing required.
• claude-mem - Plugin that captures and injects session context for cross-session memory.
  • Strengths: ~10x token savings via 3-layer workflow; SQLite + FTS5 + Chroma retrieval; multi-IDE compatibility.
  • Caveats: requires Node.js 18+; lossy memory compression; vector-DB initialization overhead.
• Everything Claude Code - Performance-optimization system and tuning reference for the Claude Code agent harness.
  • Strengths: harness performance focus; benchmark-driven tuning; concise patterns.
  • Caveats: niche perf-focused audience; assumes Claude Code familiarity.
• get-shit-done - Meta-prompting and context-engineering system for agent-distributed large projects.
  • Strengths: parallel isolated 200k-token contexts; persistent cross-session docs; auto-diagnose failures.
  • Caveats: sequential six-command workflow; best for large structured projects.
• graphify - Skill that turns code, docs, and media into navigable knowledge graphs.
  • Strengths: confidence tagging on relationships; god-node and surprising-connection detection; inline-comment rationale linking.
  • Caveats: local AST processing only; multimedia transcription via external APIs; first-run scales with project size.
• gstack - Garry Tan's exact Claude Code setup with 23 opinionated role-based tools (CEO, Designer, Eng Manager, Release Manager, Doc Engineer, QA).
  • Strengths: role-based metaphor across the dev lifecycle; mirrors a production YC workflow; broad active community.
  • Caveats: highly opinionated about the specialist-team model; less flexible for unconventional workflows.
• Karpathy-Inspired Claude Code Skills - Compact Karpathy-style coding skills distilled into Claude Code-compatible format.
  • Strengths: Karpathy minimalism applied to skills; concise, pragmatic patterns; quick to adopt.
  • Caveats: personal interpretation; small skill count; opinionated style.
• Matt Pocock's Skills - Personal Claude Code skill pack from Matt Pocock, drawn straight from his .claude directory.
  • Strengths: opinionated commands (/grill-me, /write-a-prd, /tdd, /triage-issue, /git-guardrails); failure-mode focus across planning/testing/safety; readable shell-based skills.
  • Caveats: small skill set; reflects one engineer's workflow; no built-in orchestration.
• superpowers - Agentic skills framework and methodology for Claude Code, Cursor, Codex, and Gemini CLI.
  • Strengths: TDD/RED-GREEN-REFACTOR built in; brainstorming + planning before code; Git-worktree mgmt for parallel work.
  • Caveats: philosophy-heavy; best results need cultural adoption.
• Understand-Anything - Interactive knowledge-graph explorer for codebases and documents.
  • Strengths: multi-view exploration (structural/business/semantic); guided onboarding tours; concurrent multi-agent pipeline.
  • Caveats: initial analysis pass needed; impact analysis is code-only, not runtime.

Cursor Rules

• awesome-cursor-rules-mdc - Collection of Cursor .mdc rule files covering popular stacks.
  • Strengths: modern .mdc format; broad framework coverage; actively maintained.
  • Caveats: community contributions vary in depth; framework version drift.
• awesome-cursorrules - Configuration files that customize Cursor's behavior per project.
  • Strengths: large catalog; primary community hub; widely linked from tutorials.
  • Caveats: older .cursorrules format alongside newer .mdc; rule quality varies.

MCP Servers

• awesome-mcp-servers (punkpeye) - Community catalog of Model Context Protocol servers.
  • Strengths: large index; sortable by category; community-driven additions.
  • Caveats: not curated for quality; dead links accumulate.
• awesome-mcp-servers (wong2) - Curated list of Model Context Protocol servers.
  • Strengths: tighter curation; categorized; updated regularly.
  • Caveats: smaller than alternatives; subjective inclusion bar.
• best-of-mcp-servers - Weekly-updated ranked list of MCP servers across the ecosystem.
  • Strengths: 400+ ranked MCP servers; quality scoring; weekly updates; transparent metrics.
  • Caveats: algorithmic ranking only; not a testing service; curation gaps.
• C4 Model MCP - Structurizr's official MCP server for C4 architecture diagrams with DSL validation and PlantUML/Mermaid export.
  • Strengths: official C4 reference implementation; 23 MCP tools; Docker and Java 21 deployment.
  • Caveats: requires Java 21; experimental status; HTTP stateless mode limits some features.
• Excalidraw MCP - Official Excalidraw MCP server for streaming hand-drawn diagram creation from agents.
  • Strengths: official Excalidraw team implementation; mobile-optimized streaming UI; works with Claude/ChatGPT/VS Code/Goose.
  • Caveats: requires MCP-compa