Cca Audit

CCA-Audit

**A multi-agent code auditor for [Claude Code](https://docs.anthropic.com/en/docs/claude-code) and Codex in which no unverified finding is allowed into the fix plan.**

Eleven specialised auditors read your diff in parallel. Their findings are deduplicated, then each one is **re-derived against the real code** — mechanically, by `pyright`, `clippy`, `semgrep`, `pytest` or `hypothesis` wherever a tool can settle the claim, and by adversarial review where none can. Only verified findings that survive that gate are eligible to be fixed. The fix is then checked for scope creep, and the whole change is gated on a mapping proving every confirmed finding has a fix and every edit has a finding.

**Python and Rust** have deterministic settlers. Each language gets the claim vocabulary that actually carries information in it — Rust does not get `nullability`, because the code compiled; it gets `panic_path`, `overflow` and `error_swallow` instead. Any other language, and any claim type a backend does not declare, escalates rather than being settled by a tool built for something else.

The design constra