- Home
- Skills
- Bobmatnyc
- Claude Mpm Skills
- Requesting Code Review
requesting-code-review_skill
- Python
13
GitHub Stars
2
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill bobmatnyc/claude-mpm-skills --skill requesting-code-review- metadata.json627 B
- SKILL.md3.2 KB
Overview
This skill dispatches a code-reviewer subagent to inspect recent changes and validate implementation against the plan or requirements before continuing development. It helps catch regressions, design drift, and requirement mismatches early so fixes are small and inexpensive. Use it as a routine gate after tasks, features, or before merging to main.
How this skill works
You provide the subagent with the base and head git SHAs, a brief description of what was implemented, and the plan or requirements it should satisfy. The code-reviewer compares changes between SHAs, evaluates them against the provided plan, and returns findings with severity labels (Critical, Important, Minor) and actionable recommendations. Triage the results: fix critical issues immediately, resolve important issues before the next major task, and track minor items.
When to use it
- After completing each subagent-driven task
- After finishing a major feature or batch of tasks
- Before merging any change into main or a release branch
- When you’re stuck or need a fresh perspective
- Before a significant refactor to establish a baseline
Best practices
- Always include BASE_SHA and HEAD_SHA and a concise {WHAT_WAS_IMPLEMENTED} context
- Attach the {PLAN_OR_REQUIREMENTS} so the reviewer can verify intent vs implementation
- Treat Critical findings as blockers and fix them before proceeding
- Respond to reviewers with evidence if you disagree and request clarification rather than ignoring feedback
- Run review after fixes to verify issues are resolved
Example use cases
- Automated review after a CI job comparing HEAD to origin/main to catch regressions before merge
- Review after a multi-task batch to ensure the combined changes meet the planned feature spec
- Request review when a junior developer finishes a task to provide mentoring and catch hidden issues
- Run review prior to a large refactor to ensure the pre-refactor baseline is acceptable
- Use review when fixing a complex bug to validate the fix and look for related regressions
FAQ
Provide BASE_SHA, HEAD_SHA, a short {WHAT_WAS_IMPLEMENTED}, and the {PLAN_OR_REQUIREMENTS}. A one-line {DESCRIPTION} helps focus the review.
How should I act on different severities?
Treat Critical issues as immediate blockers. Fix Important issues before the next major task. Log Minor issues for later unless they accumulate into larger risk.