Awesome Agentic Coding

A curated list of open-source tools built specifically for agentic coding.

Agentic coding turns AI from autocomplete into an active software teammate that can plan, code, review, test, coordinate, and operate across your development workflow. This list is intentionally narrow: only tools with a clear, direct agentic-coding use case and a clear open-source license belong here.

Inspired by the structure and polish of Awesome Claude Agents, but focused only on open-source agentic coding tools.

📚 Table of Contents

• 🛠️ Agentic Coding Tools
• 🤝 Contributing

🛠️ Agentic Coding Tools

• agent-orchestrator - An agentic orchestrator for parallel coding agents that plans tasks, spawns workers, and helps automate CI fixes, merge conflicts, and code review loops.
• agent-rules - A shared rules-and-knowledge repository for coding with agents such as Claude Code and Cursor.
• agentic_coding_flywheel_setup - A bootstrapper that turns a fresh Ubuntu VPS into a fully configured multi-agent coding environment.
• agents - An open-source multi-agent orchestration toolkit for Claude Code workflows and intelligent automation.
• Aider - A terminal-based coding agent that makes repo-aware edits, works cleanly with git, and supports iterative multi-file implementation workflows.
• AutoGen - Microsoft's multi-agent framework for building conversational and event-driven agent systems with strong support for orchestration, tooling, and human-in-the-loop workflows.
• beads_rust - A local-first issue tracker for git repos that helps coding agents coordinate work without invasive project changes.
• beads_viewer - A graph-aware TUI for the Beads issue tracker, useful for agent task prioritization and dependency-aware planning.
• Claude Code MCP - A one-shot MCP server that lets Claude Code act as an agent inside another agent workflow.
• claude_code_agent_farm - An orchestration framework for running many Claude Code agents in parallel with lock-based coordination and tmux monitoring.
• Codex (OpenAI) - OpenAI's open-source lightweight coding agent that runs in the terminal and supports interactive agentic development workflows.
• codex (steipete) - A lightweight open-source coding agent built to run directly in the terminal.
• CodexBar - A lightweight status tool for keeping Codex and Claude Code token or usage limits visible while you work.
• coding_agent_account_manager - A fast account-switching utility for AI coding CLIs when teams need to manage usage caps and provider limits.
• coding_agent_session_search - A TUI and CLI for indexing and searching local session history across many coding-agent tools.
• coding_agent_usage_tracker - A unified CLI for tracking usage, remaining quota, rate limits, and costs across multiple coding-agent providers.
• CrewAI - A popular framework for assembling role-based autonomous agents that collaborate on structured tasks, making it approachable for production-oriented multi-agent apps.
• cross_agent_session_resumer - A cross-provider session portability tool that helps resume work between Codex, Claude, Gemini, and related coding agents.
• DeepCode - An open agentic coding project focused on tasks such as Paper2Code, Text2Web, and Text2Backend generation.
• destructive_command_guard - A guardrail tool for blocking dangerous git and shell commands before coding agents can execute them.
• emdash - An open-source agentic development environment built for running multiple coding agents in parallel.
• franken_agent_detection - A Rust utility for detecting installed coding-agent connectors, useful when building agent-aware automation and orchestration layers.
• frankenterm - A terminal hypervisor for AI agent swarms that captures pane state and exposes a JSON API for coordinating large coding-agent fleets.
• Goose - Block's open-source local coding agent focused on practical developer workflows, extensibility, and programmable assistant behavior.
• Kilo Code - An open-source agentic engineering platform centered on coding-agent workflows for building, shipping, and iterating quickly.
• LangChain - A broad application framework for connecting LLMs to tools, retrieval, and workflows, with deep ecosystem coverage for agentic application development.
• LangGraph - LangChain's graph-based runtime for durable, stateful, and inspectable agent workflows where control flow and recovery matter.
• LlamaIndex - A data-centric framework for building agents over private knowledge, structured data, and retrieval pipelines with strong indexing primitives.
• llm.codes - A tool for transforming developer documentation into cleaner Markdown that is easier for coding agents to consume.
• macOS Automator MCP - A macOS automation MCP server for AppleScript and JXA, useful when coding agents need local desktop control.
• mcp-agentify - An orchestrator that converts MCP servers into agents, helping compose richer agentic workflows.
• mcp_agent_mail - A coordination layer for coding agents with messaging, file leases, and audit trails.
• mcp_agent_mail_rust - A Rust implementation of coding-agent mailboxes that combines Git-backed archives, SQLite indexing, and interactive coordination workflows.
• meta_skill - A skill-management platform for AI coding agents with semantic search and local persistence.
• mux - A desktop app for isolated, parallel agentic development across multiple concurrent coding workflows.
• NTM - A tmux orchestration tool for spawning, tiling, and coordinating multiple AI coding agents.
• OpenCode - A terminal-first open-source coding agent built specifically for interactive software development workflows.
• Peekaboo - A macOS screenshot and GUI-automation tool built for AI agents, available as both CLI and optional MCP server.
• pi_agent_rust - A Rust-native AI coding agent CLI aimed at high performance and systems-level reliability.
• post_compact_reminder - A Claude Code hook that detects context compaction and reminds the agent to re-read AGENTS.md so critical rules are not lost.
• PydanticAI - A type-safe Python agent framework that uses Pydantic models to make tool calls, structured outputs, and validation far more reliable.
• rano - A network observer for tracking outbound connections from AI coding CLIs and attributing traffic to specific providers in real time.
• remote_compilation_helper - A build-routing helper that transparently offloads cargo and gcc compilation from coding agents to remote workers.
• repo_updater - A bulk git-maintenance CLI for keeping large sets of repositories in sync during multi-agent coding workflows.
• RepoBar - A repo-status tool for surfacing CI, pull requests, issues, and releases while agents work across many repositories.
• slb - A two-person-rule CLI that requires peer review and approval before agents run risky or destructive actions.
• smolagents - Hugging Face's lightweight agent framework designed for simplicity, hackability, and tight control over how tools and code execution are used.
• storage_ballast_helper - A protective daemon for AI coding workloads that anticipates disk pressure and prevents disk-full incidents before they disrupt agents.
• superpowers - An agentic skills framework and software-development methodology for working more effectively with coding agents.
• SWE-agent - A software engineering agent that takes GitHub issues and attempts to fix them automatically using an LLM-driven coding loop.
• Terminator - A terminal babysitter for killing hung tabs and jobs, useful when unattended coding-agent runs get stuck.
• tmuxwatch - A lightweight TUI for watching tmux sessions, useful when supervising terminal-based coding agents.
• ultimate_bug_scanner - A large pattern-based bug scanner designed to run as a coding-agent quality guardrail.
• vibe_cockpit - A real-time monitoring dashboard for AI coding fleets, focused on session health and output visibility.
• VibeMeter - A cost tracker for AI coding tools such as Cursor and OpenAI-backed workflows.
• VibeTunnel - A remote-control layer for commanding terminal agents from any browser when you are away from your machine.
• VoltAgent - A TypeScript agent framework focused on real-world production workflows, with emphasis on agents, memory, workflows, and server integrations.
• XC Sentinel - An Xcode build helper that produces AI-friendly output and smoother automation loops for Apple-platform coding agents.

🤝 Contributing

Contributions are welcome and encouraged. If you want to add a tool, improve a description, or suggest a new category:

1. Fork this repository.
2. Add your entry using the exact format below.
3. Make sure the tool is directly related to agentic coding.
4. Make sure the project is open source with a clear license. Licenses such as MIT, Apache-2.0, AGPL-3.0, and CC0-1.0 are acceptable when they clearly cover the project.
5. Use the official project URL and keep the description concise, specific, and useful.
6. Open a pull request with a short explanation of why the addition belongs here.

* [Tool Name](URL) - A concise, 1-2 sentence description highlighting its specific value proposition.

Please prefer high-signal additions over list bloat. Proprietary products, IDE wrappers without a clear agentic-coding focus, generic AI tooling, unrelated prompt collections, and projects without a clear open-source license should stay out. CC0-1.0 is acceptable when it clearly applies to the project.