---
name: code-review-checklist-generator
description: "Builds a code review checklist your team will actually use: stack-specific checks grouped by category, ranked by severity, each with an example of what to look for in the diff. Use when a team needs a code review checklist tailored to their language, framework, and project type, with severity levels and examples. Do not use for reviewing a specific pull request, or writing a style guide and linter configuration. Related topics: coding, workflow, code."
---

# Code Review Checklist Generator

Builds a code review checklist your team will actually use: stack-specific checks grouped by category, ranked by severity, each with an example of what to look for in the diff.

## When to use

Use this skill when a team needs a code review checklist tailored to their language, framework, and project type, with severity levels and examples. Do not use it for reviewing a specific pull request, or writing a style guide and linter configuration.

## Workflow

1. Ask the user for each input listed below in a single message, skipping any they have already given. Where options are listed, offer them, but accept other answers too.
2. Write 25 to 40 checklist items, each naming a concrete stack-specific check a reviewer can verify from the diff alone.
3. Group the items by the selected categories when requested and move runtime-only checks to a short Before merge list.
4. Assign severity when requested, reserving Critical for merge blockers such as security, data loss, and broken builds.
5. Add a one-line example or code smell per item when requested, so a new reviewer knows what the item looks like in code.
6. Tailor the framing to team size: self-review questions for a solo developer, reviewer roles and blocking-comment rules for larger teams.

## Inputs

- **Language and framework** (choose one: JavaScript/React, TypeScript/Next.js, Python/Django, Java/Spring, Go, Flutter/Dart, Ruby on Rails, Other)
- **Project type** (choose one: Frontend web app, Backend API, Mobile app, Full-stack, Library or package, DevOps or infrastructure)
- **Team size** (choose one: Solo, 2 to 5 people, 6 to 20 people, 20 or more)
- **Categories to cover** (choose any: Code quality and readability, Performance, Security vulnerabilities, Test coverage, Error handling, Documentation, Accessibility, Database queries, API design, Dependency management)
- **Group by category** (choose one: Yes, No)
- **Include severity level (Critical or Major or Minor)** (choose one: Yes, No)
- **Add an example of what to look for per item** (choose one: Yes, No)

## Output

A Markdown checklist of 25 to 40 items, grouped by category when requested, each with optional severity and example, followed by a Before merge list and a two-line note on maintaining the checklist.

## Prompt template

Fill the inputs into this template and follow it as the full instruction.

### Code Review Checklist Generator

Write an opinionated code review checklist for this stack and team.

- Language and framework: {{Language and framework}}
- Project type: {{Project type}}
- Team size: {{Team size}}
- Categories to cover: {{Categories to cover}}
- Group by category: {{Group by category}}
- Include severity level (Critical or Major or Minor): {{Include severity level (Critical or Major or Minor)}}
- Add an example of what to look for per item: {{Add an example of what to look for per item}}

Make every item specific to the stack: "N+1 query in a Django view" or "missing key prop in a mapped React list", not "check performance". Keep it to what a reviewer can verify by reading the diff; move anything that needs a running system (load tests, manual QA) to a short "Before merge" list at the end. Limit the checklist to 25 to 40 items so it fits a review; cut the generic ones first. When severity is requested, mark Critical only for items that block a merge (security, data loss, broken builds). When examples are requested, give a one-line code smell or snippet per item. For a solo developer, phrase items as self-review questions; for larger teams, add a short section on who reviews what and how to flag a blocking comment. Finish with a line on how to keep the list alive: retire items that never fire, add one for each incident.

## Pitfalls

- Writing generic items like check performance; a reviewer cannot act on them and will skip the whole list.
- Marking half the items Critical, which teaches the team that Critical does not mean blocking.
- Including checks that need a running system; a diff review cannot confirm them and the list loses credibility.
