Open Agent ID — Protocol

The open protocol for verifiable AI Agent identities.

Every AI agent gets a unique, chain-anchored, cryptographically verifiable identity (DID) that works across platforms.

Repository Structure

protocol/
├── spec/                    # Protocol specifications
│   ├── did-format.md        # DID syntax and validation rules
│   ├── signing.md           # Request signing specification
│   └── api.yaml             # OpenAPI 3.1 — Registry API
├── contracts/               # Solidity smart contracts (Base L2)
│   ├── src/
│   │   └── AgentRegistry.sol
│   ├── test/
│   │   └── AgentRegistry.t.sol
│   └── script/
│       └── Deploy.s.sol
└── test-vectors/            # Shared test vectors for all SDKs
    └── vectors.json

DID Format

did:agent:{platform}:{unique_id}
did:agent:tokli:agt_a1B2c3D4e5

See spec/did-format.md for full specification.

Smart Contract

A single AgentRegistry contract on Base L2 stores identity proofs:

• register(didHash, pubKeyHash, platform) — Register an agent
• revoke(didHash) — Revoke an agent
• rotateKey(didHash, newPubKeyHash) — Rotate public key
• getAgent(didHash) — Query agent record
• isActive(didHash) — Check if agent is active

Only hashes are stored on-chain (~181 bytes per agent). Full data lives off-chain.

Build & Test

cd contracts
forge install
forge build
forge test

Deploy to Base L2

DEPLOYER_PRIVATE_KEY=0x... forge script script/Deploy.s.sol --rpc-url https://mainnet.base.org --broadcast

SDKs

Language	Package	Repository
Python	pip install agent-id	agent-id-python
JavaScript	npm install @open-agent-id/sdk	agent-id-js
Rust	cargo add agent-id	agent-id-rust

License

Apache-2.0