bmad-code-review

Reviews code diffs adversarially using parallel review layers with structured triage and patching.

Updated Jul 9, 2026
One-click install
npx skills add https://github.com/unbakedcosmos/thumbnailer --skill bmad-code-review-unbakedcosmos
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: bmad-code-review
Source: https://github.com/unbakedcosmos/thumbnailer/tree/main/.claude/skills/bmad-code-review
Command: npx skills add https://github.com/unbakedcosmos/thumbnailer --skill bmad-code-review-unbakedcosmos

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Code reviews often miss subtle bugs, edge cases, and spec deviations because a single reviewer pass lacks diverse perspectives. This Skill runs adversarial multi-layer reviews on git diffs and turns raw findings into triaged, actionable outcomes. ## Core Features & Use Cases - Parallel adversarial review layers: Launches a Blind Hunter, Edge Case Hunter, and Acceptance Auditor (when a spec file is provided) against the same diff, then merges and deduplicates their findings. - Structured triage and severity routing: Reads surrounding source code before rating severity, then routes each finding into decision-needed, patch, defer, or dismiss buckets. - Story and sprint integration: Writes findings into story files, applies patches on request, and syncs story status back to sprint-status.yaml. - Use Case: After finishing a feature branch, ask for a review of the branch diff against main; the Skill gathers the diff, runs the review layers, presents triaged findings, and applies approved fixes. ## Quick Start Ask the assistant to run a code review on your current uncommitted changes or a branch diff against main.

Frequently Asked Questions about bmad-code-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run an adversarial code review on a git diff?▼

Provide a diff source such as uncommitted changes, a branch diff, a commit range, or a PR reference, and the workflow gathers the diff automatically. It then launches parallel review layers and presents triaged findings with severity ratings.

What review layers does this code review workflow use?▼

It uses three layers: a Blind Hunter for general adversarial review, an Edge Case Hunter for boundary and failure conditions, and an Acceptance Auditor that checks the diff against a spec file's acceptance criteria when one is provided.

Can I review code without a spec or story file?▼

Yes, the workflow supports a no-spec mode where the Acceptance Auditor is skipped. Findings that would require a decision are reclassified as patch or defer since there is no spec to resolve ambiguity against.

What happens if a review subagent fails or returns nothing?▼

The failed layer is recorded in a failed_layers list and the review continues with the remaining layers. If all findings end up empty while layers failed, the workflow warns that the review may be incomplete rather than declaring it clean.

Does the code review update sprint status automatically?▼

Yes, when a story key is available it updates the story file status to done or in-progress based on review outcomes and syncs the matching entry in sprint-status.yaml. If no story key exists, it notes that sprint sync was skipped.

What are the limitations of automated diff review?▼

Very large diffs over roughly 3000 lines should be chunked by file group to stay effective. Severity ratings also depend on reading surrounding source code, so findings from diffs without accessible call sites may be less precise.