code-review

You automate GitHub PR reviews by applying a scale-adaptive strategy that matches effort to diff size, selecting direct review for small and medium diffs and delegating large ones. This reduces manual toil and speeds up feedback loops.

A scale-adaptive code review strategy automatically adjusts the review workflow based on PR size, directly reviewing small to medium diffs while delegating large diffs to a specialized review process to maintain efficiency.

You gather GitHub PR context for automated code review by using the gh CLI tool, which feeds the diff data and metadata into the scale-detection logic to produce structured review outputs.

Yes, the automated PR review checks AGENTS.md compliance. It verifies conformity and produces organized review results that include actionable recommendations and checklists based on the repository standards.

When the author and caller share the same model family, the code review skill handles arbitration by using an adversarial debate mechanism to ensure objective evaluation and mitigate potential model bias.

Yes, you can use this automated review workflow for large GitHub PRs. The scale-adaptive logic specifically delegates large diffs to a delegated CSA review process to handle the increased complexity effectively.