Skip to content

Security: QWED-AI/.github

Security

SECURITY.md

Security Policy

Supported Versions

We support security updates for the latest major version of all QWED packages.

Project Supported Version
qwed-verification ✅ Latest v2.x
qwed-finance ✅ Latest v1.x
qwed-legal ✅ Latest v0.x
qwed-ucp ✅ Latest v1.x
qwed-mcp ✅ Latest v1.x

Reporting a Vulnerability

Do NOT report security vulnerabilities via public GitHub issues.

If you discover a security vulnerability in any QWED project:

  1. Email: Send details to security@qwedai.com.
  2. Encryption: You may use our PGP key (ID: 0xQWEDSEC) if needed.
  3. Response: We will acknowledge your report within 48 hours.
  4. Bounty: We offer bug bounties for critical vulnerabilities on a case-by-case basis.

Critical Vulnerabilities Include:

  • Incorrect verification (False Negatives where a guard says "Safe" but it is not)
  • Sandbox escapes in code execution guards
  • PII leakage in logging

Security Best Practices for Users

  • Local Execution: QWED is designed to run locally. Avoid exposing guard endpoints directly to the public internet without authentication.
  • Input Sanitization: While QWED verifies outputs, always sanitize inputs to your notification systems.

There aren’t any published security advisories