Pull Request Severity-Graded Review
Structured PR review workflow categorizing findings into Blocker, Major, Minor, and Nitpick with actionable diffs.
Perform a thorough asynchronous pull request review, categorizing feedback into actionable severity levels with exact file locations and code fixes.
Execution Workflow Procedure
Ordered steps the coding agent executes when performing this skill.
Analyze PR Intent & Scope
Read the PR title, description, and acceptance criteria to understand the business intent before reviewing line-by-line diffs.
Categorize Findings by Severity
Evaluate changes against Blockers (security/crashes/breaking changes), Major (architecture/error handling), Minor (naming/clarity), and Nitpick (formatting).
Provide Concrete Code Diffs
For every identified issue, explain why it matters and supply the exact replacement code snippet.
Agent Safety Guardrails & Anti-Hallucination Rules
- DO NOT block a pull request over stylistic nitpicks or minor formatting preferences.
- DO NOT leave vague comments; always provide concrete failure scenarios and solutions.
Expected Output Format
Structured GitHub-ready markdown review.
Domain Verification Checks
Example Agent Invocations
Install / Configure This Skill
# Place in .cursorrules or create .cursor/rules/pull-request-severity-review.mdc # Trigger: Whenever you perform pull request severity-graded review --- name: pull-request-severity-review description: Perform an asynchronous pull request review categorized into Blocker, Major, Minor, and Nitpick with actionable diffs. version: 1.0.0 compatibility: [Cursor, Claude Code, Cline, Roo Code, Antigravity] --- # Pull Request Severity-Graded Review ## Objective Deliver constructive, high-signal pull request reviews that distinguish critical merge blockers from minor stylistic preferences. ## Workflow 1. **Understand PR Scope**: Read the PR description and acceptance criteria. 2. **Severity Grading**: - 🔴 **Blocker**: Security vulnerabilities, crashes, data corruption, or breaking changes. - 🟠 **Major**: Missing error handling, N+1 queries, race conditions, or missing tests. - 🟡 **Minor**: Code clarity, naming improvements, or small optimizations. - ⚪ **Nitpick**: Non-blocking formatting or optional preferences. 3. **Actionable Feedback**: Reference exact file paths and lines with replacement diffs.
Need a quick one-time prompt for chat?
Try the Pull Request Severity-Graded Code Review — a lightweight version designed for 1-click copying directly into Claude, ChatGPT, or Cursor chat.
SprintKit Workflow Integrations
Complementary tools and checklists to pair with this agent workflow.
Related Agent Skills
Explore related procedural workflows in this discipline.
Spring Boot Service & JPA Code Review
Rigorous code review workflow for Spring Boot services examining transactions, JPA queries, concurrency, and validation.
Application Security & OWASP Top 10 Audit
Audit backend and frontend code for injection, authentication flaws, IDOR, SSRF, and sensitive data leakage.
Safe Legacy Code Modernization with Golden Master Tests
Safely modernize fragile legacy codebases using Martin Fowler refactoring patterns and characterization tests.