Agents

Permissions for
Agents

Give AI agents explicit, revocable permissions that survive audits.

Use cases

*

Agent permissions

Model what each agent can do in verifiable policy terms.

*

Role updates

Adjust capabilities by appending new upsert/revoke events.

*

Explainability

Use decision explain output for operational debugging.

How it works

Get started in three steps

Treat agent access as signed state, not hidden application flags.

1

Model

Define agent relationship type and roles.

2

Publish

Sign and append events to your feed.

3

Enforce

Check decisions in app/API paths.

Ed25519
Signing
JWS
Verification
HTTPS
Transport
100%
Open Source