Won't another bot just add noise?
Code Terrier is precision-first: speculative nits are not reported at all, clean changes get no comment, and findings are capped and trimmed by severity. The goal is a signal you can act on.
Code TerrierCode Terrier for Teams shipping agent-written code
Built for Teams shipping agent-written code
Coding agents (Claude Code, Cursor, Codex, Copilot, Windsurf, Gemini CLI, Devin, Aider) write more code than anyone can read. Code Terrier reviews every pull request through a verified pipeline; its architecture supports up to three model families, while the current shared pool uses one configured model.
In auto mode, it re-checks later pushes in auto mode after a fix and quietly closes findings that no longer apply. The bot never approves or auto-merges — it only posts COMMENT reviews.
What you get
- The review architecture supports up to three model families and merges duplicate findings; the current shared pool uses one configured model, while BYOK can use the configured reviewer limit.
- Findings land on the exact changed line, categorized as Security, Bug, Flag, or Slop.
- In auto mode, a new commit triggers the next review; prior human comments are carried forward as bounded, untrusted context.
- The verifier holds final classification, so the reviewer never grades its own work.
Objections, answered
Does it work with my coding agent's workflow?
Code Terrier is a GitHub App, so it sees every PR your agents open. Auto mode reviews new PRs and later pushes by default; on-creation and manual modes give you control.
Start reviewing
Install Code Terrier and open a pull request. It only posts COMMENT reviews; it never approves or auto-merges.