Email security@identities.ai. Please do not open a public issue for a suspected vulnerability.

Include what you have: affected version, a description of the behaviour, and a reproduction if you have one. You will get an acknowledgement, and we will tell you what we find and when a fix ships.

This package decides whether a consequential action runs. A defect that lets an action execute without a matching authorization is in scope, and so is anything that weakens the binding between a decision and the action it authorized. Concretely, report:

• a way for guard to invoke a protected handler without an explicit allow;
• two different payloads or actions that produce the same request hash;
• a payload or action that can change between being hashed and being acted on;
• a malformed or hostile Verify response that reaches the handler, or that turns replay protection off;
• anything a handler can read or modify that the signature does not cover.

• The Ratify protocol itself. Specification, canonicalization, signature or certificate-chain behaviour goes to ratify-protocol.
• The hosted Ratify Verify service. Availability, the API, or anything server-side goes to security@identities.ai with the service name in the subject. It is handled separately from this package.

If you are unsure which applies, send it to security@identities.ai and say so.

This package is pre-1.0. Fixes land on the latest released version; there are no maintained release branches yet.