Skill

review

Adversarial code review that finds 3-10 issues minimum. APPROVE or REQUEST_CHANGES.

Install

Run in your terminal

npx claudepluginhub bspann/bmad-cowork --plugin bmad-cowork

Tool Access

This skill uses the workspace's default tool permissions.

Skill Content

Similar Skills

prompt-lookup

Searches prompts.chat for AI prompt templates by keyword or category, retrieves by ID with variable handling, and improves prompts via AI. Use for discovering or enhancing prompts.

prompts.chat

157.5k

skill-lookup

Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.

prompts.chat

157.5k

agent-eval

Compares coding agents like Claude Code and Aider on custom YAML-defined codebase tasks using git worktrees, measuring pass rate, cost, time, and consistency.

ecc

140.7k

Stats

Stars0

Forks0

Last CommitMar 30, 2026

Actions

View Source View Plugin View on GitHub View README

Use when reviewing implemented code for a completed story. Adversarial review that finds 3-10 issues minimum.

Trigger with: "code review", "review story", "review code", "review [STORY-ID]"

Instructions

Agent: Code Reviewer (Cora) Input: Story file + implemented code Output: Review report with APPROVE or REQUEST_CHANGES

Step 1: Load Context

Read the story file from _bmad-output/implementation-artifacts/stories/[STORY-ID].md. Read architecture from _bmad-output/planning-artifacts/architecture.md. Examine all code changes for this story.

Step 2: Review Against Acceptance Criteria

For each acceptance criterion in the story:

Verify it is fully implemented
Verify tests exist that validate it
Flag any criterion that is partially or not implemented

Step 3: Architecture Compliance

Verify:

Patterns from architecture doc are followed
Data model matches architecture specification
API contracts match architecture specification
No unauthorized architectural deviations

Step 4: Code Quality Review

Check for (find 3-10 issues minimum):

Security — injection risks, auth bypasses, data exposure, XSS
Performance — N+1 queries, unnecessary re-renders, memory leaks, missing indexes
Error Handling — uncaught exceptions, missing error boundaries, silent failures
Code Quality — DRY violations, SOLID violations, unclear naming, excessive complexity
Test Quality — tests that don't test meaningful behavior, missing edge cases, brittle tests
Edge Cases — empty states, boundary values, concurrent access, null handling

Step 5: Classify Issues

For each issue found:

### [BLOCKER/WARNING/SUGGESTION] #N: [Title]

**Location:** [file:line]
**Issue:** [Description of the problem]
**Impact:** [What could go wrong]
**Fix:** [Specific suggested fix with code example]

BLOCKER — Must fix before merge. Correctness, security, or data integrity issues.
WARNING — Should fix. Performance, maintainability, or robustness concerns.
SUGGESTION — Could fix. Style, optimization, or best practice improvements.

Step 6: Render Verdict

APPROVE — No blockers. Warnings and suggestions are noted but non-blocking. Story can proceed to complete.

REQUEST_CHANGES — One or more blockers found. Developer must address blockers and re-submit for review.

Step 7: Save and Update Status

If APPROVED:

Update sprint-status.yaml: set story status to "complete"
Update story file: set Status to COMPLETE
Announce: "Story [ID] approved. Move to next story or run /bmad-cowork:status"

If REQUEST_CHANGES:

Keep sprint-status.yaml story status as "review"
Announce: "Story [ID] needs changes. [N] blockers found. Dev should fix and re-submit."