Code review

Skill viktorbezdek/skillstack/code-review/skills/code-review

Skills I use and develop to deliver better outcomes faster and with less effort.

Install

npx -y skills add viktorbezdek/skillstack --skill code-review

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

10 stars10 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Reviews existing code and pull requests using multi-agent swarm analysis covering security, performance, style, test coverage, and documentation quality. Extracts and prioritizes PR comments, performs security audits, and generates actionable fix plans with file:line references. Use when the user asks to review code, review a PR, audit code for security, assess code quality, analyze pull request comments, get feedback on existing code, or perform a code audit. NOT for writing new code or implementing features (use other development skills), NOT for finding and fixing runtime bugs or errors (use debugging), NOT for writing tests or setting up test infrastructure (use testing-framework), NOT for TDD methodology (use test-driven-development).

SKILL.md

5.9 KB, as published. Nobody here has run it

Comprehensive Code Review Skill

This skill is a curated merge of four complementary code review skills:

Evidence-Based Code Review: Evidence-based multi-agent swarm review with specialized reviewers
PR Comment Analysis: Extract, consolidate, and prioritize GitHub PR comments
Development Workflow Specialist: TDD, debugging, performance optimization, and automated code review
Consultant: AI-powered deep analysis for complex architectural and security reviews

When to Use

Use this skill when:

Reviewing PRs systematically with multiple quality dimensions
Extracting and consolidating feedback from multiple reviewers
Need evidence-based findings with file:line references
Assessing security vulnerabilities, performance bottlenecks, or code quality
Creating action plans from PR comments
Complex architectural decisions requiring deep analysis
Audit requirements mandate systematic review (compliance, release gates)

Do NOT use for:

Simple formatting fixes (use linter/prettier directly)
Non-code files (documentation, configuration without logic)
Trivial changes (typo fixes, comment updates)
Generated code (build artifacts, vendor dependencies)

Core Capabilities

1. Evidence-Based Code Review

Every finding MUST include:

Code location: [file:line] with surrounding context (5 lines before/after)
Evidence type: DIRECT, STYLE_RULE, or BEST_PRACTICE
Severity: CRITICAL, MAJOR, MINOR, or NIT
Confidence score: 0.0-1.0
Suggested fix with specific code changes

2. PR Comment Analysis

Fetches ALL comments from GitHub PR (inline + conversation)
Groups comments by file path and code section
Identifies "High Consensus Issues" (2+ reviewers same concern)
Three-phase analysis: Consolidation, Context Validation, Impact Analysis

3. Multi-Agent Swarm Review

Five specialized review agents working in parallel:

Security Reviewer: Vulnerabilities, unsafe patterns, secrets
Performance Analyst: Bottlenecks, optimization opportunities
Style Reviewer: Code style, best practices, maintainability
Test Specialist: Test coverage, quality, edge cases
Documentation Reviewer: Comments, API docs, README updates

4. AI-Powered Consultation

Supports 100+ LLM providers through LiteLLM
Deep analysis for complex architectural decisions
Async operation with session management

5. TRUST 5 Validation Framework

Truthfulness: Code does what it claims
Relevance: Changes are appropriate for the context
Usability: Code is maintainable and understandable
Safety: No security vulnerabilities or data risks
Timeliness: Follows current best practices

Severity Levels

CRITICAL: Architecture-level issues (security vulnerabilities, data loss risks)
MAJOR: Module-level issues (performance bottlenecks, maintainability problems)
MINOR: Function-level improvements (code style, readability)
NIT: Line-level suggestions (formatting, naming)

Validation Rules

NEVER approve code without evidence
ALWAYS provide line numbers with every finding
VALIDATE findings against multiple perspectives
DISTINGUISH symptoms from root causes
AVOID false confidence: flag uncertain findings as "needs manual review"
Findings require 2+ confirming signals before flagging as violations

See Extended Patterns for detailed finding templates, PR comment workflow, multi-agent swarm execution, CI/CD integration, and quality gate configuration.

Success Criteria

Violations Detected: All quality issues found with ZERO false negatives
False Positive Rate: <5% (95%+ findings are genuine issues)
Actionable Feedback: Every finding includes file path, line number, and fix guidance
Root Cause Identified: Issues traced to underlying causes, not just symptoms

Resources Reference

Scripts

Script	Purpose
`pr-comment-grabber.py`	Extract all PR comments from GitHub
`pr-comment-filter.py`	Filter comments by criteria
`analyze-pr.sh`	Full PR analysis workflow
`multi_agent_review.py`	Multi-agent swarm orchestration
`security_scan.sh`	Security vulnerability scanning
`performance_check.py`	Performance bottleneck detection
`style_audit.py`	Style and best practices audit
`consultant_cli.py`	AI consultation CLI

References

Reference	Description
`extended-patterns.md`	Detailed examples and workflow patterns
`analysis-prompt.md`	LLM prompt for PR comment analysis
`analysis-prompt-v2.md`	Enhanced analysis with validation
`github-api.md`	GitHub API endpoints reference
`impact-analysis-methodology.md`	Risk assessment framework
`validation-workflow.md`	Three-phase validation process
`best-practices.md`	Code review best practices
`review-categories.md`	Severity and scope categories

Templates

Template	Purpose
`review-checklist.yaml`	Code review checklist
`security-rules.json`	Security rule definitions
`performance-thresholds.json`	Performance thresholds