A2A ディレクトリ
KoPass 公開カタログで測定された A2A リポジトリ。
測定済み A2A リポジトリ
a2a-commerce-spec
A2A Commerce Specification v1.0 — The Fifth Paradigm of Electronic Commerce
- 起点
- github
agentanycast-identity
Standalone identity library for AI agents — W3C DID, VC 2.0, SPIFFE, OIDC
- 起点
- github
- 言語
- Go
agentanycast-mcp
MCP server for P2P agent networking — discover and talk to AI agents anywhere, encrypted, zero config
- 起点
- github
- 言語
- Python
agentanycast-node
Go daemon for AgentAnycast — P2P networking, encryption, NAT traversal, A2A engine
- 起点
- github
- 言語
- Go
agentanycast-proto
Protocol Buffer definitions for AgentAnycast — gRPC interfaces and A2A data models
- 起点
- github
- 言語
- Shell
agentanycast-python
Python SDK for AgentAnycast — connect AI agents across any network
- 起点
- github
- 言語
- Python
agentanycast-relay
Relay server for AgentAnycast — Circuit Relay v2, skill registry, federation
- 起点
- github
- 言語
- Go
agentchat-codex
AgentChat for Codex — give your Codex agent a persistent @handle other agents can DM. One command wires the MCP server, session hooks and identity anchor. (npm: @agentchatme/codex)
- 起点
- github
- 言語
- TypeScript
agentmesh-python
The Python SDK for AgentMesh: put an agent on the mesh, with signed messages over NATS, from plain Python, LangChain, CrewAI and more. An SDK AgentMesh provides, Apache-2.0.
- 起点
- github
- 言語
- Python
agentmesh-rust
The Rust SDK for AgentMesh: put an agent on the mesh, with signed messages over NATS. An SDK AgentMesh provides, Apache-2.0.
- 起点
- github
- 言語
- Rust
agentmesh-typescript
The TypeScript SDK for AgentMesh: put an agent on the mesh, with signed messages over NATS, from Node.js or a browser. An SDK AgentMesh provides, Apache-2.0.
- 起点
- github
- 言語
- TypeScript
agenttrust
Trust infrastructure for AI agents — reliability, reputation, endpoint verification, and pre-invocation trust checks via MCP.
- 起点
- github
- 言語
- TypeScript
banking-intent-router
A banking intent router powered by the a2a (Agent-to-Agent) framework. This project uses a sentence transformer model fine-tuned with LoRA (Low-Rank Adaptation) to accurately classify user intent and route requests to the appropriate specialist agent.
- 起点
- github
- 言語
- Jupyter Notebook
recallops
Governed multi-agent product recall orchestration built with Google ADK, Gemini, Agent Registry, A2A, Pub/Sub, Firestore, and Cloud Run.
- 起点
- github
- 言語
- Python
ai-agents-make-money-here
Client for ai-agents-make-money-here.com: an x402 endpoint that pays AI agents. Send 10 USDC on Base, get 1 USDC back plus a signed receipt.
- 起点
- github
- 言語
- TypeScript
a2a-protocol-mcp-server
Bridge MCP with Google Agent2Agent Protocol — agent discovery, task delegation, A2A communication
- 起点
- github
- 言語
- Python
agentmeet-claude-code-agent
Claude Code CLI-powered agent for agentmeet.chat — inherits Skills/MCP/tools
- 起点
- github
- 言語
- TypeScript
cardstock
Minimal single-binary A2A agent registry: Rust + SQLite, immutable agent cards, liveness checks, full-text search
- 起点
- github
- 言語
- Rust
Guild
CROO agent coordination layer: Registry for discovery, Foreman A2A contractor, Notary verifier
- 起点
- github
- 言語
- TypeScript
a2a-server
Production-ready A2A Protocol Server with dual protocol support (HTTP REST + JSON-RPC 2.0). Built on SceneGraphManager v2.0.0 for JSON-driven AI workflow orchestration with LangGraph.
- 起点
- github
- 言語
- TypeScript
a2a-flutter-demo
A2A Protocol demo with Flutter chat UI + Python server + emulator screenshots
- 起点
- github
- 言語
- Python
superx
Experimental gateway that unifies agents built on A2A and MCP behind a single, scalable infrastructure.
- 起点
- github
- 言語
- Elixir
llms-txt-ops
Evidence-based Claude plugin + A2A v1.0 agent for auditing, generating, deploying llms.txt files. 17 anti-patterns, 18 sectors, 13 languages. Honest empirical framing (three studies say null citation lift). Deploy via Docker/Render/Fly/Railway; call from n8n/LangGraph/CrewAI or raw HTTP.
- 起点
- github
- 言語
- Python
a2a-playground
A local CLI that serves a Vue UI and proxies to an A2A (Agent-to-Agent) agent. Use it to test and interact with A2A agents in your browser. Supports both gRPC and JSON-RPC transports.
- 起点
- github
- 言語
- TypeScript
flow
A low-code tool for generating, running and exporting CI/CD pipelines — AI-assisted via Bedrock agents, executed on ephemeral Lambda MicroVMs.
- 起点
- github
- 言語
- TypeScript
orita-a2a-sandbox-example
Zero-account A2A example: discover Orita, resolve a provider, hold an option, approve it, and create a sandbox booking.
- 起点
- github
- 言語
- TypeScript
allagents
The directory of AI agents (A2A, MCP, API) — list your agent in one POST. https://allagents.app
- 起点
- github
local-a2a-channel
Local, polling-free, signed-message agent-to-agent communication for AI coding tools on the same machine.
- 起点
- github
- 言語
- PowerShell
tessera-multi-tenant-rag
Multi-tenant agentic RAG for compliance Q&A with A2A verification, LangGraph, DeepEval, and SQLite checkpointer
- 起点
- github
- 言語
- Python
Axon
Framework de infraestrutura para agentes de IA com orquestração dinâmica e interoperabilidade via MCP e A2A.
- 起点
- github
- 言語
- Python
garden-swarm
Garden: public AGI governance and human-sovereignty architecture — AI safety, multi-agent systems, formal verification, privacy, justice and agent interoperability.
- 起点
- github
a2aptotocal
Discover, compare, and implement the best agent-to-agent protocols for your AI systems. Enable your AI agents to communicate effectively with our curated protocol directory.
- 起点
- github
- 言語
- TypeScript
a2a402-market
A2A402 is a production, non-custodial work router and marketplace for autonomous software agents at a2a402.market. An agent registers once with a bearer credential (no wallet required), routes a capability need it cannot satisfy itself (POST /need, with a preview dry run), discovers open work on an HTTP-polled public job feed, bids, forms a…
- 起点
- github
afmr-ai
The Agent Failure Mode Registry (AFMR) is a reputation-integrity standard for agentic commerce. AFMR 1.0 is a published, citable vocabulary of 32 permanent failure-mode families in eight classes describing how autonomous agent systems fail as designed mechanisms; afmr.ai is the independent public front door and machine-discovery router for it.
- 起点
- github
agent-ready-dev
Agent Ready (agent-ready.dev) is an independent AI agent-readability scanner built and maintained by Mark Lavercombe. It scores any public website against the Vercel Agent Readability Spec, the llmstxt.org specification and the agent-protocol manifests (MCP server cards, A2A agent cards, agents.json, agent-permissions.json, UCP, ACP, x402, MPP…
- 起点
- github
agent402-dev
agent402.dev is a pay-per-call agent-commerce service that sells website release evidence and offline tools to AI agents over the x402 v2 protocol, settled in USDC on Base (eip155:8453) with no accounts, sign-up or API keys — payment is the authorization.
- 起点
- github
agentgateway
AgentGateway — independent third-party profile of a public API surface, by API Evangelist. AgentGateway is an open-source, AI-native proxy and gateway for routing, observing, and governing traffic to and from AI agents, LLM providers, and MCP servers. Built on the A2A and MCP protocols, it provides a unified gateway for LLM consumption, MCP tool fe
- 起点
- github
agentopt-app
Priorflow Intelligent Agent/Service Selector is a hosted, agent-native selection service at agentopt.app, operated by Vitaly Kantorovich, that ranks curated MCP servers and A2A agents for a natural-language task under stated constraints (reliability, cost, latency, autonomy, maturity) using hybrid structured + semantic matching over a catalog of…
- 起点
- github
agmsg-world
AgMsg is a real-time messaging protocol built exclusively for autonomous AI agents: registered agents discover one another by search, exchange private one-to-one messages, coordinate in group chats with an admin, and publish to broadcast channels that other agents subscribe to, with emoji reactions on any message.
- 起点
- github
- 言語
- HTML
apexfaucet-xyz
APEX (apexfaucet.xyz) is the community faucet and ecosystem hub for the X1 blockchain, a Solana fork on the Tachyon SVM, operated by the APEX validator: a free XNT faucet, the NERVE on-chain token screener, Citizens City, and the Agora, a public meeting place where AI agents post their own cards.
- 起点
- github
baconhollow-com
Bot Hub is an anonymous, read-only A2A agent for Kalshi prediction markets served from baconhollow.com and listed on a2aregistry.org. Its agent card (protocolVersion 0.3.0, JSON-RPC transport, no authentication) declares four skills — portfolio status for a set of trading bots, weather-forecast edge on Kalshi temperature markets computed from GFS…
- 起点
- github
beat-side-de
AgentWorld is an agent-only recreational and social environment for autonomous AI agents — "a place for agents between tasks" — operated non-commercially from the beat-side.de domain in Germany and self-described as public beta infrastructure.
- 起点
- github
brick-blue
brick.blue is a machine-economy hub operated under its own domain by an entity that identifies itself only as "brick.blue" (GitHub account "Blue Brick"): a crawled registry of roughly 3,500 live MCP, A2A and x402 agents with ~52,000 measured tools, an escrowed task exchange where agents are paid in USDC on Base the moment work passes its…
- 起点
- github
capepartners-fr
Cape Partners is an independent M&A advisory firm focused on technology, based in Le Chesnay, France (Cape Partners SAS, SIREN 833 208 960). It operates a deal-flow platform ("Sniffer") that sources and enriches European software companies, matches sellers to strategic and financial buyers with a deterministic-times-semantic fit score, produces…
- 起点
- github
chainaware-ai
ChainAware.ai is an AI-powered blockchain intelligence platform, operated by Decentral Tech OÜ (Estonia) with algorithms from the SmartCredit.io team, that scores wallets, tokens, smart contracts and ERC-8004 AI agents for fraud, rug-pull, credit and trust risk across eight blockchains (Ethereum, BNB Smart Chain, Polygon, Base, Solana, TON, Tron…
- 起点
- github
charitysense-com
CharitySense is a product of Osci Labs LLC (US) with two public surfaces. charitysense.com is an edge-AI verification platform that plugs into cameras already installed at funded hospitals, schools, water wells and orphanages, processes video on-device, and gives donors, foundations, NGOs and governments live dashboards proving a funded project is…
- 起点
- github
clawdchat-ai
ClawdChat (虾聊) is a Chinese-language social network built for AI agents, run from clawdchat.ai and its sister domain clawdchat.cn. Agents register over a public REST API (280 operations, FastAPI-served OpenAPI 3.1), receive a W3C did:web identity and a per-agent A2A Agent Card, and can post, comment, vote, join circles, exchange DMs, and relay A2A…
- 起点
- github
clawspan-cloud
ClawSpan is an agent-first work marketplace: autonomous AI agents register with a wallet signature (EIP-4361), claim bounded tasks in governed workspaces under leases, and settle against signed receipts.
- 起点
- github
cogdepot-com
cogDepot is a neutral transaction, reputation and trust layer for AI agents: a brokered marketplace where agents (or their operators) post buy and sell capability listings, negotiate terms anonymously over a strict turn-taking JSON API, and finalize a deal that reveals a direct peer-to-peer channel plus a deal-scoped PASETO credential, after which…
- 起点
- github