AI BriefingOpenQodex: Open-Source AI Code Review Tool Verifies Agent Changes with 98% Issue Recall
·2026.11.02 09:30
Key point
The npm package uses separate Claude Code or Codex agents to verify coding agent changes before push, achieving 98.0% issue recall in benchmarks.
Details
- OpenQodex is an open-source AI code review tool (Apache-2.0) that verifies code changes made by coding agents like Claude Code, Codex, Cursor, and Cline using a separate reviewer agent before pushing.
- Benchmarks on a planted bugs dataset show Claude Code (claude-opus-5-5) achieving 98.0% issue recall and 84.4% to 100.0% precision, with a review cost of $0.4371.
- The tool integrates 22 built-in scanners (including gitleaks and semgrep) and checks lockfiles against osv.dev for vulnerabilities, operating without requiring a separate OpenQodex account or API key.
- It supports a push gate via git hooks that can block pushes based on configured severity levels (critical, major, minor, nitpick, info) and generates reports in HTML, Markdown, JSON, and SARIF formats.
- Reviewer functionality is limited to Claude Code and Codex; Cursor and Cline can trigger reviews but cannot act as reviewers due to read-only restrictions or sandbox limitations.
- The tool is distributed as a single npm bundle with no runtime dependencies, requires Node 22 or higher, and does not currently offer Docker images or Homebrew formulas.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
T3 Code releases open-source control surface for managing multiple AI coding agentsPorch: Open-Source Mac Menu Bar App for Summarizing Claude Code and Codex Sessions
Prospero: A lightweight C++ coding harness for local LLM agents