Summary

This project provides production-grade engineering skills for AI coding agents, addressing the tendency for AI to bypass critical software development practices. It targets engineers and power users by enabling AI agents to consistently apply senior-level workflows, quality gates, and best practices across the entire software lifecycle, thereby enhancing code reliability and quality.

How It Works

The core approach uses structured SKILL.md files as executable workflows for AI agents, covering six development phases: Define, Plan, Build, Verify, Review, and Ship. These skills are activated via seven slash commands or automatically based on context. The methodology emphasizes "process, not prose," embedding hard-won engineering judgment and principles (e.g., Hyrum's Law, test pyramid, Shift Left) directly into step-by-step agent actions. Each skill includes anti-rationalization tables and non-negotiable verification steps, ensuring disciplined execution beyond typical AI shortcuts.

Quick Start & Requirements

• Primary Install: For Claude, use /plugin marketplace add addyosmani/agent-skills followed by /plugin install agent-skills@addy-agent-skills. For local development, clone the repository and run claude --plugin-dir /path/to/agent-skills.
• Prerequisites: A compatible AI agent environment (e.g., Claude) and Git. Skills are plain Markdown, adaptable to any agent supporting system prompts or instruction files.
• Integration: Seamlessly integrates with Cursor (.cursor/rules/), Windsurf (rules configuration), GitHub Copilot (.github/copilot-instructions.md), and other agents.
• Documentation: Setup guides are available for Cursor, Windsurf, Copilot, and general usage (docs/).

Highlighted Details

• Features 19 distinct skills, each a detailed workflow with frontmatter, process steps, rationalizations, and verification criteria.
• Provides 7 slash commands that map directly to the software development lifecycle phases.
• Includes 3 specialist agent personas (code-reviewer, test-engineer, security-auditor) for targeted reviews.
• Offers 4 supplementary reference checklists for testing, security, performance, and accessibility.
• Embeds core principles from Google's engineering culture, such as the test pyramid, Hyrum's Law, and Shift Left, directly into actionable workflows.

Maintenance & Community

The provided README does not detail specific contributors, sponsorships, or community channels (e.g., Discord, Slack). Contribution guidelines are referenced via CONTRIBUTING.md.

Licensing & Compatibility

• License: MIT.
• Compatibility: The permissive MIT license allows for broad use in projects, teams, and tools, including commercial applications, without significant restrictions.

Limitations & Caveats

The effectiveness of these skills is contingent on the AI agent's capability to accurately interpret and execute structured Markdown workflows. The opinionated nature of the skills enforces specific engineering practices that may require adjustment for teams with divergent established processes. Successful integration necessitates compatibility with supported agent platforms.