
codex-security
Find code security vulnerabilities

What you get
Scan your codebase for security vulnerabilities using AI analysis. Define policies in SECURITY.md, find and validate findings with deep mode, then fix issues. Works with OpenAI, Anthropic, Google Gemini, and other providers. Publish findings with deduplication and severity classification.
Why it exists
Defines security policies and findings scope upfront through SECURITY.md, so all scans respect the same standards without per-repo configuration.
How it works
Findings service deduplicates by embedding similarity and runs independent reviews, then persists groups to SQLite for tracking across runs and repositories.
For
Teams that want AI-assisted security scanning integrated into CI/CD without building the scanning harness themselves.
- Use cases
- / Scan repositories in bulk
- / Generate component-scoped security policies
- / Deduplicate and classify findings across repos
- ai-security
- application-security
- cli
- code-scanning
- codex
- codex-security
- cybersecurity
- devsecops
- nodejs
- npm
Useful? Star it. That is the only thanks most maintainers get.
Also from openai
- codex★ 127k
- whisper★ 110k
- openai-agents-python★ 29k