code-reviewer

Automating GitHub pull request reviews requires coordinating parallel evaluations across spec, tests, security, and performance to aggregate findings into a structured merge-readiness verdict. This process uses local subagents and gh CLI access to redact secrets before posting comments.

Parallel PR evaluation runs independent checks across spec conformance, tests, architecture, security, performance, and code quality simultaneously. It aggregates findings with severity levels, suppresses duplicates, and produces a structured merge-readiness decision based on critical issues.

Yes, gh CLI access is required to run automated code reviews on pull requests. The evaluation process requires gh CLI to parse PR frontmatter, execute local subagents for multi-aspect checks, and post structured review comments back to the repository.

Security checks are handled as one dimension of parallel PR evaluation, while PII or secrets redaction is applied before posting review comments. The evaluator aggregates security findings with severity levels and suppresses duplicates to produce clear merge-readiness feedback.

Determining merge readiness involves aggregating multi-aspect PR review findings into a structured verdict with severity levels. If critical issues are found, the PR returns to a fix loop; otherwise, it is deemed ready to merge based on the standardized evaluation.