Agent Toolkit
自动化任务
2Tracecat — Agentic Security Automation
Tracecat is an open-source security automation platform for teams and AI agents, built on Temporal with sandboxed tool runs and a self-hostable UI.
Hermes Agent — Self-Improving Multi-Channel AI Agent
Hermes Agent is a self-improving CLI and gateway agent with memory, skills, cron, and multi-provider support for shared terminal and chat workflows.
评测
14Claude Cookbooks — Claude API Notebook Recipes
Claude Cookbooks is a repo of copyable notebooks for Claude API patterns (tool use, RAG, vision, evals) so you can prototype faster (42.9k★).
DSPy Micro Agent — CLI + FastAPI + Evals
evalops/dspy-micro-agent is a minimal agent runtime (CLI + FastAPI + eval harness); verified 73★ and documents `micro-agent ask` and `run_evals.py --n 50`.
Awesome Agentic Patterns — Practical Blueprints
awesome-agentic-patterns is a curated catalogue of reusable agent patterns covering planning, tool use, memory, evaluation, and safety loops.
RagaAI Catalyst — LLM Eval + Tracing SDK
RagaAI Catalyst is a Python SDK for managing LLM projects with evaluation, dataset management, trace/agentic tracing, and prompt/guardrail workflows.
AgentScope Studio — Agent Tracing UI
AgentScope Studio is a local visualization toolkit: Projects/Runs UI, OpenTelemetry-based tracing, and eval dashboards for AgentScope agents.
MiroThinker — Open Deep Research Agent with 256K Context
MiroThinker is an open deep-research agent stack with long context, tool-heavy configs, and benchmarked BrowseComp results for serious evaluation work.
Skills
59Swarmclaw — Self-Hosted Multi-Agent Runtime + MCP
Self-hosted multi-agent runtime with memory, delegation, schedules, and MCP tools. Run swarms via web dashboard or CLI. Verified 481★; pushed 2026-05-14.
Koog — Kotlin/JVM Agent Framework with MCP Tools
A JVM agent framework for predictable, fault-tolerant agents, with MCP tool integration; verified 4187★, pushed 2026-05-14.
ZenGuard — Runtime Guardrails for AI Agents
A real-time trust layer for agents with prompt-injection/PII/secrets detectors and tiered access; verified 150★, pushed 2026-02-03.
Chorus — Multi-LLM Peer Review for Code
CLI + MCP server that runs 2–4 LLM reviewers in parallel using your existing subscriptions; verified 455★, pushed 2026-05-14.
Pydantic Deep Agents — Python Deep Agent Harness
Batteries-included Python agent harness with tool-calling, memory, hooks, and a terminal assistant CLI; verified 788★, pushed 2026-05-14.
RubyLLM::Agents — Rails Engine for Production Agents
Rails engine for building and monitoring LLM agents with a DSL, retries/fallbacks, cost analytics, and a dashboard; verified 124★, pushed 2026-04-19.
Workflows
6CKA-Agent — Trojan Knowledge Attack Agent (Research)
Research code for studying trojan-knowledge attacks in agent systems, with reproducible scripts and configs; verified 203★, pushed 2026-05-13.
SafestClaw — Offline-First Assistant CLI
Install SafestClaw with pipx to run an offline-first assistant CLI (LLM optional) with commands for news, crawling, reminders, and more.
Obsidian Agent Client — Bring Agents into Obsidian
Obsidian Agent Client adds ACP-based agent integrations (Claude Code/Codex/Gemini) to Obsidian; install via BRAT or Releases and use adapters (2,016★).
Maestro — Multi-Agent Orchestration for Coding CLIs
Maestro is an Apache-2.0 multi-agent orchestrator with 39 specialists for Gemini CLI, Claude Code, Codex, and Qwen Code, plus review/debug entrypoints.
Bernstein — Audit-Grade Orchestrator for CLI Agents
Bernstein coordinates CLI coding agents in parallel worktrees with signed audit chains, deterministic scheduling, and evidence trails.
Awesome A2A — Agent2Agent Tools, Servers & Clients
Curated Agent2Agent (A2A) directory covering agents, servers, and client SDKs for building interoperable multi-agent systems.
MCP 配置
7Headroom — Context Compression + MCP for Agents
Local context optimization layer: proxy/wrap/CCR + MCP tools to compress logs/files/RAG for agents; verified 1742★, pushed 2026-05-14.
flyto-core — Replayable Agent Execution Engine
flyto-core is an engine with traced, replayable steps and MCP-native modules; verified 270★ and includes 32+ recipes to start.
haiku.rag — Agentic RAG CLI + MCP Server
haiku.rag is an agentic RAG toolkit with CLI, Python API, and MCP server; verified 524★ and supports `add-src`, `ask --cite`, and `serve --mcp`.
Archestra — Secure AI Platform (MCP Governance)
Archestra centralizes MCP servers with governance and observability; verified 3,653★ and ships a Docker quickstart for a private MCP registry.
AstrBot — Multi-IM Agent Framework + Plugins
AstrBot is an open-source agent framework integrating IM platforms, LLMs, and plugins, with a uv quickstart and Docker docs for production deployments.
Hyperterse — Agentic Server Framework (MCP + A2A)
Deploy agents and MCP tools from one service. Exposes `/mcp` Streamable HTTP and A2A routes, with DB adapters and OpenTelemetry hooks.
CLI 工具
6Kocoro — Local Agent Runtime CLI (shan)
Run agents locally via the `shan` CLI with interactive TUI, daemon mode, tools, and MCP integrations; verified 263★ and installs via npm or script.
Hyperspace AGI — P2P Agent Network CLI
hyperspaceai/agi is the Hyperspace network repo and CLI entrypoint. README shows Pods, distributed training, and a one-line installer.
Orca — Multi-Agent IDE for CLI Agents
IDE for running multiple CLI coding agents with worktree isolation, diff review, and notifications. Works with Claude Code, Codex, OpenCode, and more.
Lyrie — Autonomous Security Agent CLI + ATP SDK
Lyrie is an autonomous security agent with a Python CLI (`lyrie-omega`) plus an Agent Trust Protocol SDK (`@lyrie/atp`) for cryptographic identity.
MemPalace — Local-First AI Memory CLI
MemPalace is a local-first memory system with a `mempalace` CLI (+ MCP tools), hitting 96.6% R@5 on LongMemEval with no LLM required.
Ruflo — Swarm Orchestration CLI + MCP
Ruflo is a CLI + MCP stack for orchestrating agent swarms in Claude Code, with one-line install and an MCP server mode you can add via `claude mcp add`.
知识库
5awesome-trading-agents — Trading Agents + MCP List
Curated list of trading agents, market-data MCPs, and skills, with “If you only read three” starters and bilingual docs. Verified 114★; pushed 2026-05-11.
learn-agentic-ai — Agentic AI Patterns with Dapr + MCP
Learning repo for agentic AI at scale: Dapr/Kubernetes patterns plus MCP/A2A concepts and linked guides. Verified 4162★; pushed 2025-10-26.
MemoryGraph — Graph MCP Memory Server (pipx)
MemoryGraph is a graph-based MCP memory server for coding agents; verified 199★ and shows a pipx install + `claude mcp add` setup for Claude Code.
Memorix — Cross-Agent Memory Control Plane
AVIDS2/memorix is a local-first cross-agent memory layer (MCP + CLI/TUI); verified 443★ with stdio `memorix serve` and an HTTP dashboard mode.
Self-Evolving Agents Survey — Lifelong Systems
Awesome-Self-Evolving-Agents is a survey collection on self-evolving AI agents and lifelong systems, focusing on feedback, memory, and iteration loops.