What problem does it solve? Reviewing code changes and documents consistently is hard: reviewers miss edge cases, skip verification checks, and apply inconsistent standards. This Skill runs structured review passes — adversarial critique, edge-case tracing, verification-gap analysis, and editorial structure/prose review — and reports findings in one canonical shape. ## Core Features & Use Cases - Multi-lens review: Runs adversarial, edge-case-hunter, verification-gap, structure, and prose lenses, each with its own method and findings shape, selected automatically by content type or explicitly by the caller. - Customizable workflow: A customize.toml file with team and user overrides controls lenses, style guides, output format, standing review directives, and report destinations. - Parallel subagent execution: Independent lenses run simultaneously as subagents when available, with dependent lenses (like prose after structure) chained on prior findings. - Use Case: Ask for a review of a pull request diff; the Skill stages the diff, runs the applicable lenses, and returns a JSON findings array plus a markdown report grouped by lens with locations, fixes, and consequences. ## Quick Start Ask the assistant to review the current branch diff or a specific file, for example: review the uncommitted changes in this repository.
