Codex Skills Library

A curated library of reusable Codex skills for developers, individuals, and teams.

This repository helps you turn OpenAI Codex into a consistent, reliable AI assistant by packaging common workflows into reusable skills instead of repeating prompts.

What is this?

Codex Skills Library is a collection of focused skills for common developer workflows:

• Code generation and debugging
• Refactoring and optimization
• Testing and documentation
• System design and decision support
• Engineering operations and reliability

Each skill is small, focused, triggered by natural language, and designed for progressive disclosure.

Repository structure

Skills are organized by domain under skills/. Each skill is a folder with a SKILL.md that includes frontmatter (name, description) and a short workflow.

Skills by domain

Foundation

• codebase-orientation: Map entry points, key modules, and build/run paths; include local setup commands, env vars, and a safe starter task.
• git-basic-helper: Provide minimal, safe git commands with clear intent and warnings for destructive actions.
• debugging-checklist: Give a prioritized checklist from repro to isolation, logging, and hypothesis validation.
• error-message-explainer: Translate compiler/runtime errors into plain language, likely root causes, and targeted fixes.
• linter-fix-guide: Explain lint rules, show the expected pattern, and propose the smallest fix.
• config-file-explainer: Summarize purpose, key sections, defaults, and which settings are safe to change.
• data-structure-chooser: Recommend a structure based on operations and constraints, with time/space tradeoffs.
• dependency-install-helper: List required runtimes, install steps by platform, and verification commands.
• small-script-generator: Generate tiny automation scripts with safe defaults and a usage example.
• ticket-breakdown: Turn a request into small, testable steps with dependencies and validation checks.
• log-summarizer: Group errors, identify the first failure, and propose concrete next actions.

Docs

• readme-polish: Add missing setup details, env vars, and troubleshooting guidance while keeping it concise.
• function-docstrings: Document purpose, parameters, return values, and error conditions in project style.
• release-notes-drafter: Group changes by feature/fix/breaking, translate to user impact, and add migration notes.
• team-onboarding-guide: Provide access/setup steps, key services, and a first-week learning plan.

Testing

• unit-test-starter: Draft tests around core behavior, edge cases, and failure paths with run instructions.
• integration-test-planner: Identify integration points, scenarios, data flows, and needed fixtures/mocks.
• bug-repro-plan: Produce a minimal repro with exact steps, environment details, and expected vs actual.

API

• buffer-api: Build Buffer GraphQL API integrations for auth, channel discovery, post creation, pagination, errors, and rate limits.
• api-request-builder: Build curl/fetch requests with auth, headers, and response inspection tips.
• api-contract-checker: Compare endpoints and payloads, flag breaking changes, and suggest versioning.
• api-error-taxonomy: Standardize error codes, payload shape, and logging guidance.
• graphql-query-optimizer: Reduce query depth and N+1 patterns using batching, caching, and pagination.

Frontend

• accessibility-basic-check: Check contrast, labels, focus order, and keyboard navigation for regressions.
• css-layout-helper: Diagnose layout intent and provide minimal flex/grid fixes with rationale.
• cli-ux-improver: Improve CLI defaults, help text, and errors with actionable next steps.

Backend

• caching-strategy-helper: Choose cache type, TTLs, and invalidation triggers based on freshness needs.
• queue-processing-patterns: Define idempotency, retries, visibility timeouts, and dead-letter routing.
• feature-flag-playbook: Plan targeting, rollout phases, monitoring, and cleanup criteria.
• system-design-draft: Outline components, data flow, storage, and tradeoffs with open questions.

Data

• sql-query-starter: Build SELECT queries with filters, ordering, limits, and parameter placeholders.
• db-migration-reviewer: Check lock risk, backfills, ordering, and rollback safety.
• data-governance-check: Map data sensitivity, retention, access, and audit requirements.

Infra

• ci-failure-triage: Classify failures, identify flake sources, and propose stabilizing fixes.
• observability-setup: Define key metrics, add correlation IDs, and set alert thresholds.
• iac-reviewer: Validate security groups, IAM rules, state changes, and deletion risk.
• platform-migration-plan: Create phased migration steps with compatibility and rollback.
• multi-region-strategy: Choose active-active or active-passive with data consistency tradeoffs.
• cross-service-debugger: Correlate logs/traces across services and isolate the failing hop.
• zero-downtime-migration: Plan dual writes, cutover checks, and rollback triggers.

Security

• security-quick-scan: Check auth, input validation, secrets, and risky dependencies.
• threat-modeling: Identify assets, trust boundaries, threats, and mitigations.
• dependency-risk-audit: Review licenses, vulnerabilities, and maintainer health.
• compliance-readiness: Map controls to evidence, highlight gaps, and assign remediation.
• config-hardening: Enforce safer defaults, validation, and secret handling.

Performance

• performance-trace-guide: Capture stable traces, identify hotspots, and propose fixes.
• scalability-assessment: Estimate growth, find bottlenecks, and suggest scale strategies.
• performance-budgeting: Set budgets, thresholds, and CI or release enforcement.
• cost-optimization-review: Identify cost drivers, right-sizing, and savings estimates.

Reliability

• reliability-slo-sla: Define SLIs, error budgets, and alert thresholds tied to impact.
• incident-postmortem: Create a timeline, root cause analysis, and action items.

Architecture

• architecture-review: Evaluate assumptions, bottlenecks, and failure modes with tradeoffs.
• domain-modeling: Identify entities, invariants, and bounded contexts with interfaces.

Planning

• refactor-roadmap: Sequence refactors with tests, flags, and checkpoints.
• tech-debt-portfolio: Catalog debt, estimate impact, and prioritize with ROI.
• roadmap-prioritization: Score initiatives by impact, effort, and dependencies.
• org-standardization: Propose minimal standards and adoption enforcement.
• vendor-evaluation: Compare vendors by security, integration, cost, and risks.
• dependency-upgrade-plan: Order upgrades, note breaking changes, and plan rollback.
• simple-refactor: Improve naming and structure without behavior changes.
• pr-reviewer: Identify logic bugs, edge cases, and missing tests with severity.

Get started

1. Browse skills/ and open a SKILL.md that matches your task.
2. Trigger the skill by asking Codex for that workflow.
3. Improve or extend skills as your team learns what works best.

Step-by-step tutorial: https://proflead.dev/posts/codex-skills-explained-101/

Video tutorial: https://youtu.be/d3Ydt6LyGeY

Contribute

Contributions are welcome. Please open an issue for ideas or send a PR with new skills, improvements, or fixes. If this library is useful, please star the repo and share it with your team.