Skip to content

Security: balogun14/OpenReviewer

Security

SECURITY.md

Security Policy

OpenReview AI is security-focused software. Please report vulnerabilities responsibly.

Supported Versions

The project is pre-1.0. Security fixes are applied to the main branch until versioned releases begin.

Reporting a Vulnerability

Do not open a public issue for vulnerabilities.

Until a dedicated security email is available, report privately to the repository owner or maintainer through GitHub.

Please include:

  • affected component
  • impact
  • reproduction steps
  • relevant logs or payloads, if safe to share
  • suggested fix, if known

Scope

Security reports are especially useful for:

  • GitHub webhook signature validation
  • GitHub App authentication
  • provider API key handling
  • prompt injection risks
  • unsafe prompt includes or path traversal
  • secret leakage in logs or findings
  • repository data exposure
  • unsafe GitHub comment posting

Project Safety Rules

OpenReview AI must remain defensive by default.

  • Do not add autonomous exploitation behavior.
  • Do not add destructive testing behavior.
  • Do not add credential extraction behavior.
  • Do not send repository data to providers unless explicitly configured.
  • Do not log secrets, provider keys, webhook secrets, or raw credentials.

Security review prompts may discuss exploitability only to explain defensive impact and remediation.

There aren't any published security advisories