Best skills for: Code review
Code review isn't just about spotting syntax issues—it's a gateway to better architecture and domain alignment. The right AI coding skills can elevate your review from superficial to substantive. Instead of relying solely on linters, consider skills that enforce design principles. For instance, a Domain-Driven Design (DDD) skill helps you assess whether aggregates, entities, and bounded contexts are correctly implemented, catching domain inconsistencies early. Another powerful addition is a Definition of Done skill tailored to your platform, which acts as a living checklist that ensures every PR meets acceptance criteria, from test coverage to documentation. A common mistake is to apply too many generic rules; focus on skills that match your project's stack and practices. Use these skills to automate the checks that humans tend to overlook, then reserve your energy for strategic questions—like whether the solution fits the long-term vision or introduces unnecessary complexity.
24 skills selected
Structured Code Review
Development
Performs a systematic code review across five quality dimensions: correctness, security, style, architecture, and testing. It parses targets like staged changes, commits, ranges, or PRs, reads the full files, and produces a structured verdict with findings by severity. Best used as a gate before merging, after subagent work, or when a second opinion is needed.
Shizuku Task Runner
Development
Use task commands for building, running, linting, and testing the Shizuku application. This skill ensures consistent use of the project's build system (taskfile.dev) instead of direct Go invocations, preventing build errors and enforcing project standards.
PR Description Generator
Development
Generates thorough pull request descriptions by analyzing git diffs, commit history, and related files. Helps when creating or updating PRs, especially to produce reviewer-friendly summaries, change lists, and testing notes.
Code Review
Development
Rigorous code review of uncommitted changes covering architecture, quality, security, and best practices. Outputs a review report and actionable plan for user approval.
Skill Reviewer
Development
Reviews agent skills against a 10-rule rubric, generates P0-P3 reports, and can scaffold a compliant skill.
Code Review
Development
This skill performs quick, focused code reviews on specified files or recent git changes. It checks for bugs, security issues, performance problems, and adherence to best practices, categorizing findings by severity. It helps developers catch issues early and improve code quality efficiently.
Pull Request Summarizer
Development
Analyzes and summarizes a pull request by extracting context, changes, and impact. Useful for code reviews, onboarding to a PR, or preparing release notes.
Code Review Orchestrator
Development
Orchestrates specialized subagents to perform a comprehensive code review and produce a markdown report. Works in PR mode (with --pr) or on the current branch, analyzing changes and invoking reviewers for Go, architecture, conventions, Bubble Tea, and documentation.
Deep Code Review
Development
Deep code review agent that checks for quality, security, patterns, and architectural consistency. Use /code-review [file or branch] to review.
Principal Architect MigrateIQ
Development
Principal architect for the MigrateIQ platform. Makes architectural decisions, reviews code, plans features, enforces consistency, and guides development across the full stack: SvelteKit 5, Tailwind CSS v4, TypeScript, Drizzle ORM/PostgreSQL, MCP server, and Claude Code skills. Use when planning new features, reviewing approaches, making technology choices, refactoring, or debugging architectural issues.
Skills Validator
Development
Validates and reviews skill definitions against the Agent Skills specification. Checks YAML frontmatter, naming conventions, description quality, file structure, and best practices, providing actionable improvement suggestions with before/after examples. Ideal for catching errors and improving skill quality before committing.
Address PR review findings to zero
Development
Guides a PR author through pre-merge review findings to zero: blockers, low-confidence, and style suggestions are fixed, refuted, deferred, or explicitly accepted.
Create a pull request
Development
Creates a pull request for the current branch, including formatting, committing, rebasing, and running the pre-PR gate.
Go Code Review for fzuhelper-server
Development
Specialized Go code review skill for the fzuhelper-server microservices backend, covering Hertz, Kitex, GORM, Redis, and project conventions.
Scope Creep Detector
Development
Analyzes git diffs against a stated intent to detect scope creep, unrelated files, broad PRs, and oversized changes, then suggests keep, split, or justify actions.
PR Quality Assessment
Development
Extract implementation guidance and assess risk from pull requests before implementing fixes. Use the comprehensive checklist to evaluate code quality, testing, style, documentation, and potential breaking changes. Helps determine whether a PR is a reliable reference implementation and plan tests or edge cases.
Verify change conformance
Development
Verify that an implementation matches change artifacts (specs, tasks, design) using the openspec CLI.
QC Quality Gate
Development
Quality gate that runs five parallel AI reviewers on every code diff against main. Each agent checks a specific dimension (code review, simplification, consistency, robustness, scope) and all must pass before commit. Helps catch issues early and enforce codebase standards.
Code Review Scripts
Development
Standalone code review agent for R, Stata, Python, Julia scripts. Checks code quality, reproducibility, figures, and polish.
Recursive Language Model (RLM)
Development
Processes large inputs that exceed single-pass capacity by decomposing tasks into manageable chunks, spawning parallel Haiku sub-agents to process each chunk, and synthesizing results into a unified output. Uses a Sonnet supervisor for decomposition and synthesis, while Haiku workers handle chunk processing in parallel, reducing costs and improving speed.
MR Split Analysis
Development
Analyzes all changes on the current branch (uncommitted and committed) and recommends whether to split them into smaller, focused MRs. Uses heuristics like backend/frontend separation, migration-first, and cohesion.
Adversarial Code Reviewer
Development
Adversarial code review using three hostile personas to catch blind spots before merging.
PR Quiz
Development
Interactive quiz that tests whether a PR author genuinely understands their own branch before requesting review, covering changes, rationale, edge cases, and risks. Grounded in real code with honest grading.
OpenLark Code Standards Checker
Development
Audits code consistency in OpenLark projects: architecture, API patterns, parameter validation, naming, and exports. Outputs actionable checklists with evidence (path:line) and risk levels, ideal for PR reviews or before adding a new API.
How to choose
- How are these skills selected?
- An algorithm picks them among the published skills that match this page's criteria, such as category, tags or compatible tools. It ranks them by the quality score a language model gave them. On ranking pages, it also weighs GitHub stars and the security rating. Nobody reviews them one by one.
- What do the security ratings mean?
- A language model reads the SKILL.md file, without running it, and gives it a rating. Safe means it sees no meaningful execution risk. Caution means the skill uses powerful tools such as bash, docker or the network for a legitimate purpose. Risky means it asks for destructive actions, for secrets to be sent out, or for obfuscated code to run; a skill rated Risky is never published. The rating does not replace your own reading of the file.
- How do I install a skill?
- Open any skill page and follow its install instructions for your tool — Claude Code, Cursor or Copilot. Each skill lists the exact steps so you can get it running in a couple of minutes.