repo-review-code

Review multi-agent repositories for concurrency and state machine issues using a bounded, findings-first analysis of core runtime modules and stateful logic. This generates actionable findings with file references, severity levels, and trigger scenarios for remediation.

A findings-only code review identifies correctness, maintainability, and API misuse risks without handling security, docs, or tests. It analyzes high-risk implementation surfaces and outputs a standardized packet containing file references and severity for downstream synthesis.

Static analysis can detect implementation drift in multi-agent systems by bounding the review to manifests, entry points, and core runtime modules. It outputs a review artifact detailing specific trigger scenarios and severity levels for identified drift.

Prepare a review packet for multi-agent synthesis by generating a bounded, findings-first review of maintainability and correctness risks. The packet must include file references, severity, and trigger scenarios before handing the artifact to a synthesis specialist.

Avoid a bounded findings-first approach when you need comprehensive security audits, documentation generation, test coverage analysis, or final report synthesis, as this review strictly focuses on correctness, maintainability, and API-use risk detection.