peer-review

Multi-perspective code review automation coordinates Scout, Research, and reviewer agents through a blackboard architecture to surface diverse findings. It enforces session lifecycle management to produce consumable review outputs for feature branches and modules.

Blackboard architecture for code review coordinates independent Scout, Research, and reviewer agents to share findings on a common workspace. This enables synthesized, multi-perspective assessment of diffs and modules before merges.

Run structured code review on major refactors by triggering a session on the target codebase. The session handles progress tracking and synthesizes agent findings into a final document with ADR-style documentation.

Automated peer-review works on diffs, pull requests, feature branches, and modules requiring rigorous scrutiny. It orchestrates session creation and synthesis to output consumable review findings tailored to the target codebase changes.

Document code review findings for architecture decisions by synthesizing multi-perspective agent outputs into ADR-style documentation. The review session lifecycle automatically progresses from tracking to final document generation.

Use multi-agent code review for major refactors, feature branches, or modules requiring rigorous scrutiny where standard single-pass reviews miss diverse issues. The blackboard architecture coordinates distinct agent perspectives to surface comprehensive findings.