Security Policy

Reporting Vulnerabilities

If you discover a security vulnerability in Provable AI, please report it privately.

Please do not open public issues for security vulnerabilities.

We will investigate and respond as quickly as possible.

Provable AI is designed to protect the integrity and verifiability of AI decisions.

Key guarantees include:

• deterministic decision protocols
• cryptographic hash-linked ledgers
• ED25519 signature verification
• independent proof verification

Any modification to decision records should cause verification to fail.

We ask researchers to allow time for fixes before publicly disclosing vulnerabilities.

We appreciate responsible disclosure and security research that helps strengthen the system.