Thank you for taking the time to report a security issue in the Seal attestation specification or the reference verifier libraries in this repository.

Please do not open a public GitHub issue for anything that could be a signature-forgery, key-disclosure, or similar cryptographic flaw.

Email security@aqta.ai with:

• A description of the issue.
• Steps to reproduce, if applicable.
• The affected component (for example: the Python verifier, the TypeScript verifier, the canonicalisation rule in the spec, the reference issuer example).
• Whether you have already tested the behaviour against an issuer other than the reference implementation.

We read every report. We do not promise a response time, and we do not pay for reports: the forgery bounty closed on 12 September 2026. Valid reports are fixed, published as test vectors where they apply, and credited in the advisory if the reporter wishes.

• Signature forgery against a receipt that was produced by a conformant issuer.
• Any way to make a verifier return valid: true for a tampered receipt.
• A deviation in the canonicalisation rule of §6 that allows two semantically-equivalent payloads to produce different canonical bytes (and therefore different signatures).
• Private-key leakage paths from the reference issuer example (noting that the example is documented as non-production and not suitable for handling real signing keys).
• Denial-of-service paths in the verifiers that a malicious receipt can trigger (excessive memory allocation, unbounded loops, etc.).

• The security of the Seal managed service itself. Those reports belong on the service's own disclosure channel at https://app.aqta.ai/security.
• Vulnerabilities in the upstream cryptography (Python) or tweetnacl (JavaScript) libraries. Please report those to the upstream projects; we will upgrade the pin once a fix ships.
• Issues in third-party issuers or verifiers that claim ATTESTATION-v1 conformance. Please contact the implementer; we are happy to coordinate disclosure for cross-implementer issues.

We prefer coordinated disclosure. Once a fix is available:

• A patch release is cut for the affected package.
• A security advisory is opened on this repository's GitHub Security Advisories tab.
• Credit is given to the reporter in the advisory, unless the reporter prefers to remain anonymous.

Thank you for keeping the attestation ecosystem honest.