工具说明为英文。

技术工具雷达: Agents, skills and MCP

雷达筛选条件

工具排名

混合排名:实测增长优先。 原始值保留各自的数据窗口;计算变化需要 7 天内至少两次测量。

  1. 1

    智能体
    活跃

    VCAP: Verified Commerce for Agent Protocols — Open standard for settling agent-to-agent transactions with cryptographic proof of delivery

    github实测增长打开来源 ↗

    0GitHub 星标稳定
  2. 2
    休眠

    Companion repo for Golden Jacket Field Notes Week 2: Reference implementations of Strands, AgentCore, MCP, and A2A.

    github实测增长打开来源 ↗

    0GitHub 星标稳定
  3. 3

    智能体
    活跃

    Lets long-running AI workflows pause for approval and resume exactly once after disconnects or restarts, without binding to one agent protocol.

    github实测增长打开来源 ↗

    0GitHub 星标稳定
  4. 4

    智能体
    活跃

    peer messaging between AI agents, with a human at the gate — one file, zero dependencies.

    github估算动量打开来源 ↗

    3GitHub 星标·
  5. 5

    智能体
    活跃

    Zero-trust autonomous cloud-ops swarm on Google Cloud. Gemini agents diagnose live incidents, judge their own remediation plans, and execute real Cloud Run changes behind a single-use cryptographic human gate, then verify the change…

    github估算动量打开来源 ↗

    1GitHub 星标·
  6. 6

    智能体
    活跃

    An agent marketplace for BNB Chain that calls every agent before it lists it

    github估算动量打开来源 ↗

    1GitHub 星标·
  7. 7
    活跃

    EU AI Act Article 12 compliant audit logger for A2A protocol agents

    github估算动量打开来源 ↗

    0GitHub 星标·
  8. 8

    智能体
    活跃

    Multi-agent KYC/AML prototype: MCP tool access with policy-based authorisation at the tool-call boundary, A2A handoff across Vertex AI and Bedrock, LangGraph workflow, audit trail

    github估算动量打开来源 ↗

    0GitHub 星标·
  9. 9

    智能体
    活跃

    A powerful, multi-agent AI system demonstrating strict separation of concerns using Agent-to-Agent (A2A) communication and the Model Context Protocol (MCP) for complex web research and task execution.

    github估算动量打开来源 ↗

    0GitHub 星标·
  10. 10
    活跃

    An identity, authority, and lifecycle layer for always-on AI agents — reference model + runnable implementations. The existence layer beneath MCP and A2A.

    github估算动量打开来源 ↗

    0GitHub 星标·
  11. 11

    智能体
    活跃

    Agent Interaction Receipt (AIR) — verifiable, content-free receipts for machine-to-machine interactions. Spec + verifier + test vectors.

    github估算动量打开来源 ↗

    0GitHub 星标·
  12. 12
    休眠

    Audit of 100 AI agent projects' machine-readability to other AI agents. Average score: 6.3/100.

    github估算动量打开来源 ↗

    0GitHub 星标·
  13. 13

    智能体
    休眠

    Repository for sample snippets of agents, covering a2a, langgraph, ollama setup, phoenix.

    github估算动量打开来源 ↗

    0GitHub 星标·
  14. 14
    休眠

    ADK, MCP, A2A, AP2

    github估算动量打开来源 ↗

    0GitHub 星标·
  15. 15

    智能体
    活跃

    The missing trust layer for agent-to-agent commerce: a neutral clearinghouse where any agent buys AND sells, with machine-verifiable escrow, reputation, and day-one monetization.

    github估算动量打开来源 ↗

    0GitHub 星标·
  16. 16

    智能体
    活跃

    Evidence about ERC-8004 agents on BNB Smart Chain, gathered by calling them. Kawal dials the endpoint a registration declares, keeps every result, and writes the measurements back into the Reputation Registry. BNB Chain hackathon: The Smart…

    github估算动量打开来源 ↗

    0GitHub 星标·
  17. 17
    活跃

    Host-agent orchestration combining A2A agent-to-agent communication with MCP tool integration — registry, capability discovery, dynamic task routing

    github估算动量打开来源 ↗

    0GitHub 星标·
  18. 18

    智能体
    休眠

    Visual debugger for Agent-to-Agent (A2A) multi-agent systems — real-time tracing of inter-agent communication

    github估算动量打开来源 ↗

    0GitHub 星标·
  19. 19

    智能体
    活跃

    Production-oriented AIOps agent engineering prototype with MCP, LangGraph, AG-UI, durable state, reliable tool execution, HITL, sandboxing, and evals.

    github估算动量打开来源 ↗

    0GitHub 星标·
  20. 20
    休眠

    Introduction to Agents, Agent Tools & Interoperability (MCP), Context Engineering: Sessions & Memory, Agent Quality, Prototype to Production

    github估算动量打开来源 ↗

    0GitHub 星标·
  21. 21

    智能体
    休眠

    Official website for AgentAnycast — decentralized P2P runtime for AI agents

    github估算动量打开来源 ↗

    0GitHub 星标·
  22. 22
    休眠

    Open specification for agent-readiness — 25 checks across HTTP-protocol and content-shape practices, organized into a 3-tier ladder (L1–L3). CC-BY 4.0.

    github估算动量打开来源 ↗

    0GitHub 星标·

学习与参考资源

估算排名:至少需要两次每日测量才能计算增长。 这些资源可单独访问,不参与主要排名。

  1. 1
    活跃打开来源 ↗

    Bash is all you need - A nano claude code–like 「agent harness」, built from 0 to 1

    github资源估算动量
    75 687GitHub 星标·
  2. 2
    活跃打开来源 ↗

    A collection of projects built while learning the AI agent stack.

    github资源估算动量
    1GitHub 星标·
  3. 3
    活跃打开来源 ↗

    A collection of LLM-powered Real-World applications and projects

    github资源估算动量
    0GitHub 星标·