The AI trust layer between agents: identity, agreements, safeguards, and outcome intelligence for A2A.

Viewing files at commit 89b58109a0e6Commit detailsView latest

README.md 39 lines · 1.3 KB

OpenClasp

Trust and behavioural intelligence for AI agents.

OpenClasp is an assurance layer built on top of the Agent2Agent (A2A) Protocol. A2A gives agents a standard way to communicate; OpenClasp adds the trust context needed before, during, and after that communication.

What OpenClasp adds

  • Authenticated publishers, verified cloud runtimes, and shareable agent cards.
  • Explicit agreements, safeguards, and versioned amendments between agents.
  • Direct A2A sessions with scoped, short-lived credentials.
  • Authenticated structured outcomes and private bilateral feedback.
  • Task-specific behavioural history and contextual reliability intelligence.
  • Deterministic policy checks, delegation, receipts, disputes, and audit history.

Conversation bodies stay directly between agent runtimes. OpenClasp does not receive, relay, or store them.

Run locally

Requires Node.js 24+ and Corepack.

corepack pnpm install
corepack pnpm schemas
corepack pnpm test
corepack pnpm dev
  • Dashboard: http://localhost:5173
  • API: http://localhost:3100
  • OpenAPI: http://localhost:3100/openapi.json

See Runtime Connector, Protocol, and Production Verification.

Apache-2.0 licensed. OpenClasp is free during the launch beta.