Skip to content

Security: mattyopon/faultray-app

Security

SECURITY.md

Security Policy

Supported Versions

Version Supported
11.x (latest) Yes
10.x Critical fixes only
< 10.0 No

Reporting a Vulnerability

Do not open a public GitHub Issue for security vulnerabilities.

Please report security issues via private disclosure:

What to include

  1. Description of the vulnerability
  2. Steps to reproduce (proof-of-concept if available)
  3. Potential impact
  4. Any suggested mitigations

Response timeline

Stage Timeline
Acknowledgement Within 2 business days
Initial assessment Within 5 business days
Fix & CVE assignment (if applicable) Within 30 days for critical, 90 days for others
Public disclosure After fix is released

CVE Process

For vulnerabilities that qualify as CVEs:

  1. We will request a CVE ID via MITRE or coordinate with the reporter
  2. A GitHub Security Advisory will be published
  3. A patch release will be tagged and pushed to PyPI/npm simultaneously
  4. Credit will be given to the reporter (unless they prefer anonymity)

Known Security Controls

  • All API routes require Supabase JWT authentication
  • Row Level Security (RLS) enabled on all Supabase tables
  • Stripe webhook signatures verified on /api/webhooks/stripe
  • Content Security Policy headers set via next.config.ts
  • Rate limiting: 60 req/min per IP on simulation endpoints
  • No secrets in client-side bundles — all keys in environment variables

Dependency Monitoring

Dependencies are monitored via:

  • npm audit — run in CI on every PR
  • Dependabot alerts (planned — see LICENSE-02)

Bug Bounty

There is currently no formal monetary bug bounty program. However, responsible disclosure is actively encouraged and recognized:

  • Hall of Fame: Reporters who disclose valid vulnerabilities are credited in our CHANGELOG.md and on our website (unless anonymity is requested)
  • Priority support: Access to a dedicated support channel for researchers during the disclosure period
  • Future program: A formal monetary bug bounty program is planned for 2026. Researchers who disclose before launch will be grandfathered in

In scope: Authentication bypasses, RLS policy violations, SSRF, injection attacks, data exposure via API Out of scope: Denial of service, rate limit bypass (already mitigated), brute force on non-production systems

Report to: security@faultray.com

There aren't any published security advisories