Browse the library

Deja-vu provides persistent memory for Claude Code, Cursor, and other AI coding agents by indexing local session history, enabling instant recall of past solutions without explicit prompting.

A pi extension that brings Claude Code-style autonomous sub-agents and workflow orchestration, supporting parallel execution and mid-run steering.

An open-source AI agent harness in native Rust, offering GUI, CLI, and webapp, supporting MCP, skills, plugins, and agent teams.

This experimental system enables temporary collaboration rooms where humans and AI agents can exchange context and tasks, then disappear without requiring permanent workspaces.

Orchestrate multiple coding agents and humans in a structured pipeline where agents propose and humans verify.

This coding agent harness for Pi transforms your terminal into an agent workspace, offering orchestration, live feedback, and review guardrails for controlled development with AI.

alook

alookai

This platform gives your local AI coding agents persistent identities and shared rooms, enabling collaboration and communication with people on your team.

This plugin provides plain-language skills for AI agents like Claude Code and Cursor, translating technical output into honest session reports, clear verdicts, and actionable steps for non-engineers.

An MCP server that gives AI coding agents efficient access to up-to-date documentation for APIs, services, and libraries.

A local MCP server and CLI providing coding agents with a persistent, Git-versioned map of the entire codebase for long-term context.

This resource provides battle-tested configurations, plugins, and hooks for Claude Code, OpenAI Codex, and Cursor agents.

AgentRQ is a self-hosted, human-in-loop task manager for AI agents that uses MCP to enable real-time collaboration and task delegation via visual boards and workflows.

MassGen is an open-source multi-agent scaling system that allows AI agents to collaborate, reason, and iteratively refine complex tasks, installable as a skill for various coding agents.

This repository provides engineering practices, design patterns, and reusable agent components for transforming Claude Code into a robust, manageable, and repeatable engineering system.

A ranked list of AI agent harnesses, frameworks, and learning resources, with an MCP server to allow agents to query and recommend harnesses.

An autonomous coding agent that transforms tasks or GitHub issues into verified pull requests with signed receipts.

Provides a lightweight long-term memory framework for LLM agents, offering structured organization and adaptive retrieval of conversational and action-aware memory.

An open-source, self-hosted terminal AI assistant and Python framework, similar to Claude Code, offering tool-calling, sandboxed execution, and MCP support for building custom agents.

This is a curated, community-driven list of Model Context Protocol (MCP) servers, tools, frameworks, and clients for AI agents.

This Python framework enables developers to build custom MCP servers and tools with features like authorized tool calling and secure secret management.

This tool provides a real-time dashboard to monitor Claude Code, Cursor, and Codex agent activity, sessions, tool usage, and subagent orchestration.

This free desktop application unifies the management of AI agent skills, allowing users to write a skill once and sync it across 32 AI coding tools via symlinks.

This CLI tool helps manage and activate AI agent capabilities, including modular skills, expert workflow roles, and slash commands, for various coding assistants.

A command-line MCP client that lets you or your agent call any MCP server's tools, resources and prompts from the shell, with OAuth and persistent sessions.