bmad-code-review
bmad-code-org/bmad-method
Review code changes with multiple independent reviewers in parallel, then triage findings.
What is bmad-code-review?
Runs parallel code reviews using several independent reviewers, then consolidates and presents the results. Use this skill when you need structured code review feedback on changes or when the user explicitly requests a code review.
- Executes code reviews with multiple independent reviewers running in parallel
- Triages and consolidates review findings from all reviewers
- Presents aggregated review results in a structured format
- Supports configurable review depth (quick or thorough workflows)
- Integrates with BMad project infrastructure for consistent review execution
How to install bmad-code-review
npx skills add https://github.com/bmad-code-org/bmad-method --skill bmad-code-review- BMad project setup in the current working directory
- uv (Python package manager) installed and available
- Absolute paths to project root and skill directory
How to use bmad-code-review
- 1.Ensure you are in or have access to the project root directory
- 2.Call the skill with 'run code review' or 'review this code'
- 3.Optionally specify review depth by naming 'quick' or 'thorough'
- 4.The skill will execute the render_skill.py script with appropriate parameters
- 5.Read and follow the workflow.md instruction printed to stdout upon success
Use cases
- Review pull request changes before merging to main branch
- Get multiple perspectives on code quality and design decisions
- Conduct thorough code audits on critical or complex changes
- Quickly validate code changes with a lightweight review pass
- Consolidate feedback from distributed review processes
- Development teams using BMad-based workflows
- Code reviewers seeking parallel review coordination
- Project maintainers managing code quality gates
- Teams needing structured, multi-reviewer feedback
bmad-code-review FAQ
The skill will offer to run the bmad skill's setup process, installing BMad if needed, then retry the command.
Yes. When invoking the skill, name either 'quick' or 'thorough' and the skill will append the appropriate workflow parameter to the command.
Check the command output for errors. If uv is unavailable or another failure occurs, the skill will report the output and halt rather than running workflow source directly.
No. The skill runs without changing the current working directory; you provide the absolute path to the project root.
Full instructions (SKILL.md)
Source of truth, from bmad-code-org/bmad-method.
name: bmad-code-review description: 'Review code changes with several independent reviewers in parallel, then triage and present the findings. Use when the user says "run code review" or "review this code"'
Run the following command exactly once without changing the current working directory. Replace {project-root} with the absolute path to the project root and {skill-root} with the absolute path to this skill's directory:
uv run --no-cache "{project-root}/_bmad/scripts/render_skill.py" --project-root "{project-root}" --skill "{skill-root}"
- When the invocation names a review selection (
quickorthorough), append--set workflow.review=<value>to the command. - On success, read and follow the one absolute
workflow.mdinstruction printed to stdout. - If the script is not found, BMad is not set up here. Offer to run the
bmadskill's setup, installingbmadfirst if you do not have it (npx skills add bmad-code-org/BMAD-METHOD --skill bmad), then run the command above once more. - On any other failure (including
uvbeing unavailable), report the command output and HALT. Do not run any workflow source directly.
Related skills
More from bmad-code-org/bmad-method and the wider catalog.
bmad-correct-course
Assess sprint change impact across PRD, epics, architecture, and UX; produce a structured change proposal.
bmad-create-epics-and-stories
Break requirements into epics and user stories with detailed acceptance criteria.
bmad-customize
Author and update customization overrides for BMad agent and workflow skills.
bmad-deep-recon
Research a topic to support a decision: draft prompts, process reports, or run parallel web searches.
bmad-forge-idea
Pressure-test half-formed ideas through questioning conversations until you can act on them or drop them with confidence.
bmad-party-mode
Orchestrate lively multi-agent roundtables with distinct personas, custom parties, and group discussions.