bmad-code-review

Review code diffs with parallel adversarial layers and triage findings into actionable categories.

1|Updated May 16, 2026
One-click install
npx skills add https://github.com/enigmaicon-eng/AI-Enterprise-OS --skill bmad-code-review-enigmaicon-eng
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: bmad-code-review
Source: https://github.com/enigmaicon-eng/AI-Enterprise-OS/tree/main/examples/BMAD-METHOD/src/bmm-skills/4-implementation/bmad-code-review
Command: npx skills add https://github.com/enigmaicon-eng/AI-Enterprise-OS --skill bmad-code-review-enigmaicon-eng

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It eliminates missed bugs and unclear requirements by running adversarial, multi-perspective reviews on code diffs and converting results into actionable categories.

Core Features & Use Cases

  • Parallel adversarial review layers: Blind Hunter, Edge Case Hunter, and (when a spec is provided) Acceptance Auditor.
  • Structured workflow with gating: Gather context, review, triage into decision_needed/patch/defer/dismiss, then guide the next actions.
  • Spec-aligned acceptance validation: When a spec or story file is available, findings are checked against acceptance criteria and context docs, then persisted into the story file.

Quick Start

Use the bmad-code-review skill to review the changes you want analyzed, including or excluding spec context as appropriate for your workflow.

Frequently Asked Questions about bmad-code-review

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

FAQPage Schema
How does multi-agent code review work for analyzing git diffs?

Yes, you can validate staged and uncommitted changes against a spec by applying the Acceptance Auditor layer, which checks findings against acceptance criteria and context docs before persisting them into the story file.

How do I triage code review findings into actionable categories?

To triage code review findings, the workflow normalizes, deduplicates, and classifies adversarial review results into decision_needed, patch, defer, or dismiss categories, guiding user-gated next actions and patch-handling steps.

What's the best way to perform adversarial testing on a commit range?

Adversarial testing on a commit range is best performed by constructing a diff from the range and invoking parallel review layers to hunt for edge cases and blind spots, resulting in deterministic triage of discovered vulnerabilities.

Can I use spec validation to check pull request diffs against acceptance criteria?

Yes, you can use spec validation to check pull request diffs by running the Acceptance Auditor, which validates diff findings against provided spec context and context docs, then persists the results into the story file.

What are the limitations of automated code review when no spec is provided?

Without a provided spec, automated code review limitations include skipping the Acceptance Auditor layer, limiting the validation to blind and edge case hunting without structured acceptance criteria checking or story file persistence.