addyosmani/agent-skills

Production-grade engineering skills for AI coding agents.

What it solves

AI coding agents often lack the disciplined engineering rigor of senior developers, frequently skipping tests, ignoring performance, or jumping straight to code without a proper specification. This project provides a library of structured workflows—called "skills"—that force agents to follow production-grade engineering best practices across the entire software development lifecycle.

How it works

The project consists of 24 Markdown-based skills that act as system prompts or instruction sets for AI agents. Each skill is designed as a process rather than a reference document, featuring:

  • Step-by-step workflows: Clear sequences of actions with checkpoints and exit criteria.
  • Anti-rationalization tables: Pre-defined rebuttals to common excuses agents make to skip quality steps (e.g., "I'll add tests later").
  • Verification gates: Strict requirements for evidence (like passing tests or runtime data) before a task is considered complete.
  • Slash commands: Pre-defined triggers (e.g., /spec, /plan, /build) that map to specific lifecycle phases.

These skills can be installed into various AI agents (such as Claude Code, Cursor, Codex, and Copilot) via a CLI or by adding the Markdown files to the agent's configuration.

Who it’s for

Developers using AI coding agents who want to ensure their AI-generated code meets professional standards for quality, security, and maintainability.

Highlights

  • Comprehensive Lifecycle Coverage: Includes 24 skills covering everything from initial requirement gathering (/spec) to deployment (/ship).
  • Agent Personas: Includes specialist personas like a Senior Staff Engineer for code reviews and a Security Engineer for vulnerability audits.
  • Broad Compatibility: Works with a wide range of AI tools including Claude Code, Cursor, Windsurf, and GitHub Copilot.
  • Autonomous Mode: The /build auto command allows agents to plan and implement tasks autonomously while maintaining test-driven verification.

Related

  • Project
    hellowind777/helloagents

    A workflow layer for AI coding CLIs that adds specialized skills, project knowledge bases, and verification gates to ensure agents complete tasks reliably.

  • Project
    FrancyJGLisboa/agent-skill-creator

    A platform that turns expert human knowledge and business evidence into tested, reusable, and governed AI agent skills for organizational use.

  • Project
    pedrohcgs/claude-code-my-workflow

    A ready-to-fork foundation for AI-assisted academic work that uses specialized agents and quality gates to automate the creation and review of papers, slides, and data analysis.

  • Project
    coleam00/skills

    A collection of modular Markdown-based skills for AI coding agents that replace bloated system prompts with task-specific procedures for planning, implementing, and validating software.

  • Project
    MengTo/Skills

    A curated library of procedural agent skills and prompt playbooks designed to help AI coding agents build high-quality UIs, games, and frontend systems consistently.