Browse the library

This resource explains agentic AI systems with runnable, CI-checked patterns, covering foundations, workflows, and implementation details for building multi-agent architectures.

Chinese translations of 7000+ skills aggregated from Anthropic, OpenAI, Vercel, Hugging Face, Obsidian and other popular skill repos, with tutorials.

Metis is a coding agent that searches, remembers, executes, and verifies code across the terminal and desktop, boosting AI/LLM coding performance.

This workflow enables AI coding agents to manage development from an idea or codebase to reviewed, tested code using an RFC-driven process and markdown files.

This tool acts as a lightweight proxy, routing Claude Code's Anthropic API calls to free or cheaper LLM providers like NVIDIA NIM or OpenRouter.

OrchestKit is a complete AI development toolkit for Claude Code, offering 106 skills, 36 agents, and 171 hooks for agent orchestration and building complex AI-driven applications.

This plugin provides a structured engineering harness for AI coding agents like Claude Code, adding planning, audit gates, and role-based workflow management to make agent output more reliable.

This Claude Code skill allows users to pit 100 sub-agents with different strategies against each other on the same task, identifying the most robust solution by having them attack and defend their work.

Han is a suite of AI skills and agents for Claude Code that provides specialists for evidence-based planning, test-driven implementation, deep code review, and documentation maintenance, streamlining product engineering workflows.

This collection offers six Claude Code plugins to manage project context, apply conditional rules, track tasks, and provide guided setup, improving the agent's efficiency and consistency across sessions.

This system orchestrates specialized AI agents for Claude Code across various development tasks, enhancing workflows with coordination, resilience, and organizational memory.

This plugin configures Claude Code into a full engineering team with 12 specialized agents and 15 automation hooks, guided by a disciplined development methodology.

Nix-based tool to declaratively pin, select, bundle and install Agent Skills for different agents via Home Manager, with skills-only plugin export.

A collection of Claude Code plugins, including an RPI (research-plan-implement) loop for structured development and avoiding hallucinations.

A Russian course and guide for Claude Code, including a starter kit with skills and subagents for advanced configuration and workflow management.

A localized collection of over 132 Cursor AI coding rules in Chinese, covering 32 technical domains for various development tasks.

Provides persistent memory for Claude Code, allowing the agent to retain context and knowledge across sessions using an external database.

Provides open skills for Claude Code and Codex to manage a personal company using AI agents, organizing tasks, memory, and workflows.

Implements a team mode for Codex with specialized subagents (Explorer, Executor, Reviewer) to delegate and parallelize complex development tasks under a main agent.

Curates a community-driven catalog of Claude Code mods (function hooks), detailing their read/write/run/network capabilities.

Provides self-hosted remote approval and live monitoring for Claude Code/Codex agents via Telegram, Feishu, or a web terminal, from any device.

Implements a governance framework for AI coding agents, enforcing a gated workflow and verifying completion evidence before shipping code.

Claude Code plugin that shows a live CO2 estimate in the status line from your token usage and saves session history locally in SQLite.

Offers a collection of 103 ready-to-use AI agent skills in SKILL.md format for various platforms like Claude Code and Cursor.