A curated collection of tools, methodologies, research, and resources for software developers in the age of GenAI.

Note: This is not a beginner's guide to AI tools. If you're already familiar with the basics (ChatGPT, Copilot, Cursor, etc.), this collection goes deeper: cool tools, extensions, open-source alternatives, videos, research papers, and interesting links to level up your AI-powered development workflow.

Frameworks and libraries for building AI agents with planning, delegation, and autonomous capabilities.

• CrewAI - Multi-agent platform with role-based agent design for collaborative AI workflows (GitHub)
• Deep Agents - Open-source agent framework with planning, sub-agents, and MCP integration (GitHub)
• HumanLayer SDK - API and SDK for human-in-the-loop approval workflows in AI agent systems (GitHub)
• LangGraph - Framework for building stateful, multi-actor AI agents with cyclic workflows and human-in-the-loop support (GitHub)
• Mem0 - Intelligent memory layer enabling AI agents to remember user preferences and learn continuously (GitHub)

AI-powered integrated development environments and code editors with autonomous coding capabilities.

• Google Antigravity - Agentic IDE powered by Gemini 3 that autonomously executes multi-step coding tasks (Website)

Central platforms for AI model discovery, sharing, and collaboration.

• Hugging Face - AI community platform with 1M+ models and Transformers library for local model execution (Website)
• Teamwork Graph - Atlassian's unified data layer with GraphQL API for AI-driven insights across apps (Website)

Tools for running AI models locally, self-hosted solutions, and privacy-focused alternatives.

• LM Studio - Desktop app for running LLMs locally with OpenAI-compatible API and developer SDKs (Website)
• Ollama - Run large language models locally with a simple CLI and API (Website)
• vLLM - High-performance LLM inference and serving library with PagedAttention and OpenAI-compatible API (GitHub)

Model Context Protocol servers, integrations, and extensions for Claude and other AI assistants.

• Context7 - MCP server delivering up-to-date, version-specific library documentation directly into LLM prompts (GitHub)
• DevSkills - MCP server for sharing AI agent skills across coding tools like Claude Code, Cursor, and Copilot (GitHub)
• Kiro Powers - Dynamic MCP tool bundles with framework expertise that activate contextually (Website)

Tools for monitoring, tracing, and debugging LLM applications and AI agents.

• LangSmith - Observability platform for debugging, monitoring, and understanding LLM application behavior (Website)

Structured methodologies and frameworks for AI-assisted software development.

• 12-Factor Agents - Design principles for building reliable, production-ready LLM-powered applications (GitHub)
• Agent OS - Structured specifications and workflows that transform AI coding agents into productive developers (GitHub)
• BMad Method - AI-driven agile framework with specialized agents for development workflows (GitHub)
• Long-Running Agent Harness - Methodology for AI agents to work across multiple context windows with progress tracking (Article)
• Spec Kit - Spec-driven development toolkit where specifications become executable implementations (GitHub)
• Tomcat AI-Enabled - Reference patterns for making 726K+ LOC codebases AI-navigable with specialized agents and MCP integration (GitHub)

Tools for creating, testing, managing, and optimizing prompts.

• Awesome AI System Prompts - Curated collection of system prompts from leading AI tools like ChatGPT, Claude, Cursor, and more (GitHub)

Tools, extensions, and utilities specifically designed for Claude Code.

• Beads - Git-backed distributed graph issue tracker providing persistent memory and task dependencies for AI agents (GitHub)
• Claude Code Plugins - Official plugin system for extending Claude Code with commands, agents, skills, and MCP servers (Docs)
• Ralph Loop - Self-referential AI loop enabling Claude Code to autonomously iterate on tasks until completion (GitHub)
• Superpowers - Skills framework enabling structured TDD workflows and extended autonomous development sessions (GitHub)
• Claude Code Sandboxing - Built-in filesystem and network isolation for safer autonomous AI agent execution (Docs)
• Claude Code Tresor - Collection of 141 professional-grade agents, skills, and commands for Claude Code (GitHub)
• Claude Trace - Logging and visualization tool that records all Claude Code interactions (GitHub)
• CodeLayer - Open-source IDE built on Claude Code for orchestrating AI coding agents in complex codebases (Website)
• Happy Coder - Mobile and web client for remote access to Claude Code with end-to-end encryption (GitHub)
• Hugging Face Skills - LLM fine-tuning through natural language commands in coding agents (Blog)
• Super Claude Kit - Persistence layer that adds cross-session context memory to Claude Code (GitHub)
• TACHES CC Resources - 27 commands, 7 skills, and 3 agents for structured AI-assisted development workflows (GitHub)
• The Definitive Guide to Claude Code - Comprehensive practitioner's guide covering installation, the Explore-Plan-Code-Commit workflow, context engineering with CLAUDE.md, and production patterns. (JP Caparas)

Tools for AI-assisted documentation generation, knowledge management, and learning resources.

• NotebookLM - Google's AI research assistant that transforms documents into summaries, podcasts, and mind maps (Website)

Open-source implementations and alternatives to popular commercial AI tools.

• PageLM - Open-source NotebookLM alternative that transforms study materials into interactive learning experiences (GitHub)

Videos, talks, and resources about AI mindset, ROI measurement, and successful AI adoption in software engineering.

• Can you prove AI ROI in Software Eng? - Stanford study with 120k+ developers shows why identical AI tools deliver ~0% lift in some orgs and 25%+ in others. Includes ROI playbook and metrics framework. (Yegor Denisov-Blanch, Stanford)
• Does AI Actually Boost Developer Productivity? - Stanford study of 100k developers reveals ~20% average productivity boost, but high variance: effectiveness depends on task complexity, codebase maturity, language popularity, and codebase size. (Yegor Denisov-Blanch, Stanford)
• No Vibes Allowed: Solving Hard Problems in Complex Codebases - How to use AI coding agents effectively in 300k+ LOC production codebases using "frequent intentional compaction" and context engineering techniques. (Dex Horthy, HumanLayer)
• AI and Junior Developers: Why the Job Will Change But Not Disappear - Historical analysis comparing AI's impact on junior developers to Engels' Pause during the Industrial Revolution, with practical advice for engineers at all levels. (Charles Anthony Browne)
• Shipping at Inference-Speed - Practitioner insights on managing 3-8 concurrent AI-assisted projects, multi-machine workflows, and why "most software does not require hard thinking." (Peter Steinberger)

Cutting-edge tools, research projects, and experimental AI applications.

• METR Task Horizons - Research measuring AI agent task completion time horizons with open-source evaluation infrastructure (Website)
• NoLiMa: Long-Context Evaluation Beyond Literal Matching - Benchmark revealing that 11 of 13 LLMs claiming 128K+ token support drop below 50% performance at 32K tokens when semantic understanding is required instead of literal matching.

Use the /add-tool [url] command to add new tools to this list. The command will analyze the tool, create documentation, and suggest appropriate categorization.

Detailed documentation for each tool is available in the tools/ directory.