opencode-review

An iterative dual-review pipeline automates code review by running uncommitted changes through both OpenCode and Claude to catch bugs, security issues, and style violations before a git commit.

A dual-model code review pipeline combines OpenCode and Claude to iteratively validate source code. It enforces a refinement loop until both models report a clean status, catching bugs and style violations that a single model might miss.

Yes, you need the OpenCode CLI tool installed and accessible in your system path. The pipeline relies on the OpenCode CLI for external model invocation to successfully execute the dual-review validation process.

To start an iterative code review on uncommitted changes, run the review command to trigger the automated pipeline. The system applies both OpenCode and Claude to your current working changes, looping until all detected issues are resolved.

Use a dual-model AI review when committing complex features that require high-assurance code quality. The dual-review process addresses single-model inconsistency by ensuring code is thoroughly vetted by two distinct AI perspectives.

If OpenCode or Claude finds bugs during the iterative review loop, the pipeline automatically enforces a refinement cycle. This loop continues until both reviewers report a clean status, ensuring all security issues and logic gaps are fully resolved.