Best skills for Codex
Codex is OpenAI's agentic coding tool — you run it from the command line and it plans, edits files and executes tasks across your project. Skills are what make those runs dependable on a specific job. A skill is a reusable SKILL.md instruction set; with Codex you give it to the agent as its working instructions, so it follows the same vetted steps every time — automating a browser, reviewing a spec, generating a test suite or shipping a refactor.
This page ranks the best Codex skills available right now. We recompute the ranking continuously from three signals: an automated quality score of each skill, its GitHub star count, and a security review that pushes risky-permission skills down and out of the top. Safe, well-documented, widely-starred skills come first. No single signal wins on its own: a heavily-starred skill with a thin or unsafe SKILL.md still ranks below a well-built, security-checked one.
The top picks are led by development skills — refactoring, code review, automation — followed by productivity and data & AI, with testing, documentation and DevOps entries too. Each skill page below explains what the skill does, its difficulty level, and how to run it with Codex.
Most of these skills are portable: the same SKILL.md usually drops straight into Claude Code, Cursor or Windsurf — check the platform badges on each card. Because we re-score the catalogue continuously as fresh skills are pulled and vetted, today's top pick may not hold next month. Bookmark this page if you want the current answer rather than a stale one.
24 skills ranked
React Hook Form Best Practices
Development
Master React Hook Form: avoid anti-patterns, manage subscriptions, and handle reset, dirty state, and controlled inputs correctly.
Emotion to Mantine Migration
Development
Migrate Emotion styled-components to Mantine components with style props and CSS modules. Use when converting .styled.tsx files or removing @emotion imports.
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.
Firecrawl Build Search
Development
Integrate Firecrawl /search into product code and agent workflows for web discovery before extraction. Use when features start with a query instead of a URL.
CCXT for TypeScript and JavaScript
Development
Comprehensive guide to using CCXT in TypeScript/JavaScript for crypto exchange integration, covering REST and WebSocket APIs, authentication, and examples.
Create Interactive Maps with kepler.gl
Data & AI
Create interactive map visualizations and export standalone HTML using the keplergl Python package. Use to plot geospatial data from DataFrames, GeoJSON, or CSV.
Configure a Chat Channel with SecretRefs
DevOps
Set up chat channels using non-interactive commands, with secrets only as SecretRefs. Covers gather, mutate, repair, and prove steps for reliable delivery.
DeFi AMM Security
Security
Security checklist for Solidity AMM contracts, liquidity pools, and swap flows. Covers reentrancy, donation attacks, oracle manipulation, slippage, and integer math.
Set up pre-commit hooks with Husky, lint-staged, Prettier
Development
Set up Husky, lint-staged, and Prettier to add a pre-commit hook that formats files, type-checks, and runs tests.
Pyrefly Type Coverage
Development
Migrate a file to use stricter Pyrefly type checking by adding annotations to all functions, classes, and attributes.
PDF Reader
Productivity
Read and extract text from PDF files, including documents, reports, contracts, and spreadsheets. Works with local files, URLs, and WhatsApp attachments. Use whenever PDF content is needed, even if not explicitly requested.
Santa Method
Testing
Multi-agent adversarial verification with two independent reviewers. Output is checked twice and iteratively fixed until both pass before shipping.
PostgreSQL Patterns
Development
PostgreSQL best practices for schema design, indexing, query optimization, and security. Includes RLS policies, upserts, and anti-pattern detection.
Playwright E2E Testing Patterns
Testing
Playwright E2E testing patterns covering Page Object Model, configuration, CI/CD integration, artifact management, and flaky test strategies.
Team Agent Orchestration
Development
Run team-based orchestration for agent squads using work items, ownership, agent Kanban, merge gates, and control pane handoffs.
Browser Use Cloud Reference
Development
Documentation reference for Browser Use Cloud, the hosted API and SDK for browser automation. Covers REST APIs v2/v3/v4, authentication, sessions, profiles, proxies, and integration patterns.
Dev Toolbar Review Guide
Security
Safety rules for the dev toolbar, PostHog client, and feature flags. Use when reviewing changes to dev-tools and related files, covering environment guards, override cookies, and SSE stream safety.
Testing Strategy for Supabase Studio
Testing
Testing strategy for Supabase Studio: writing tests, extracting logic from components, choosing between unit, component, and E2E tests.
Studio Queries & Mutations (React Query)
Development
React Query conventions for data fetching in Supabase Studio: query keys, queryOptions, mutation hooks, and imperative fetching.
CUDA Index Width in PyTorch
Development
Choose 32-bit vs 64-bit index math in PyTorch CUDA kernels. Fix large-tensor indexing overflows using canUse32BitIndexMath and related utilities.
Home Assistant Integration Knowledge
Development
Complete reference for building, testing and reviewing Home Assistant integrations, covering best practices, entity platforms, and quality scale rules.
Grafana Panel & Viz Testing Strategy
Testing
Write unit and E2E tests for Grafana visualization panels and viz utilities following repo conventions. Covers builders, canvas draw-call snapshots, and anti-flake rules.
Git Commit Guidelines
Development
Format and create git commits by following repository conventions. Review repo state, stage only relevant files, and keep commits focused.
Agent-Driven Parameter Optimization
Data & AI
Guide for optimizing configurable systems with a measurable objective, using bounded trials and evidence-based recommendations.
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.