Hi HN! I built a GitHub Action that automatically scans AI/LLM endpoints for security vulnerabilities on every push/PR.
Why? Most teams ship AI features without security testing. This action catches prompt injection, jailbreaks, and data leakage before they hit production.
How it works:
- Add 5 lines of YAML to your workflow
- Scans run automatically on push/PR
- PRs get blocked if critical vulns are found
- Full report with remediation steps
Free tier: 5 scans/month. 650+ attack vectors.
Built this because I do AI red teaming professionally (OSCP+, C-AI/MLPen). Happy to answer questions!
