multi-tool-branch-review

To run a multi-model git branch review, execute the review skill against your current branch and a chosen base branch. It orchestrates parallel Codex and Claude review passes on the git diff and consolidates findings into a single prioritized remediation plan.

A multi-model branch review audits a git diff using multiple independent AI reviewers to check for maintainability issues from different angles. It deduplicates overlapping findings from Codex and Claude, turning disagreements into one actionable remediation plan.

Yes, you can use Codex and Claude together for pre-merge code auditing. The review process launches independent review passes across both models, validates their outputs, and applies failure recovery for ACP and MCP edge cases before consolidating results.

Automated branch review handles empty reports and MCP failures through built-in failure recovery mechanisms. It validates review outputs, detects suspect or empty reports, and applies known fixes for ACP, MCP, and timeout issues to ensure a complete remediation plan.

Yes, you need python3 and git installed along with codex, claude, and Compozy dependencies. These tools are required to execute the scripted pass selection, process the git diff, and orchestrate the background execution of the review passes.

The best way to consolidate AI code review findings is to use an orchestration tool that runs multiple reviewers in parallel and deduplicates their outputs. This approach processes the git diff, validates reports, and outputs a single prioritized remediation plan.