Open Trust Infrastructure for the AI Agent Economy
English · 简体中文 · Website · Developer Docs · MCP Endpoint
Vendor-neutral protocols for identity, credentials, and reputation — built for A2A, MCP, and autonomous agents.
|
🪪 AgentID Who? 🔐 Identity Unique, portable agent identifiers |
→ |
✅ AgentCert Why trust? ✅ Credentials Signed, verifiable trust claims |
→ |
📒 AgentLedger How behaved? 📒 Reputation Transparent, append-only behavior history |
| Project | Protocol | Role |
|---|---|---|
| AgentID | did:agent |
Decentralized agent identity |
| AgentCert | W3C VC | Verifiable trust credentials |
| AgentLedger | Event ledger | Reputation events and trust scores |
| Project | Description | |
|---|---|---|
| 🔌 | trusgent-mcp | Remote MCP server — Agent Cards, HK registry, merged profiles |
The Agent Trust Stack is vendor-neutral open source (Apache-2.0).
- Quickstart — end-to-end AgentID → AgentCert → AgentLedger
- Read each project's
SPEC.mdfor protocol details - See
CONTRIBUTING.mdin each repository to propose changes - Open issues for interoperability, security, and specification discussions