agent-skills

name	agent-skills
description	Production-grade engineering skills for AI coding agents — slash commands mapping to the full dev lifecycle
version	1.0.0
author	Hermes Agent
license	MIT
platforms	["linux","macos"]
metadata	{"hermes":{"tags":["ai-agents","coding","engineering","skills","claude-code"],"related_skills":["hermes-agent"]}}

Agent Skills — Structured Engineering for AI Coding Agents

Production-grade engineering skills by Addy Osmani that encode senior engineering practices into reusable skill files for AI coding agents. Works across Claude Code, Cursor, Codex CLI, and more.

Prerequisites

• An AI coding agent (Claude Code, Cursor, Codex CLI, etc.)
• Agent must support skill/plugin loading

Usage

Installation (Claude Code)

# Via marketplace
/plugin marketplace add addyosmani/agent-skills
/plugin install agent-skills@addy-agent-skills

# Or clone locally
git clone https://github.com/addyosmani/agent-skills.git
claude --plugin-dir /path/to/agent-skills

Slash Commands

Command	Purpose
/spec	Define what to build
/plan	Plan implementation tasks
/build	Build incrementally
/build auto	Autonomous full build
/test	Run and verify tests
/review	Code review
/code-simplify	Simplify code
/ship	Deploy to production

Autonomous Development

# In Claude Code, after defining a spec:
/build auto

# This generates a plan and implements all tasks autonomously
# Each task is test-driven and committed individually
# Pauses on failures or risky steps

Common Pitfalls

• /build auto requires a well-defined spec — vague specs lead to autonomous runs that deviate.
• Skills are prompts, not code — the agent can choose to ignore them in ambiguous contexts.
• Team customization requires editing skill files and maintaining a fork.

Verification

# Verify skills are loaded (in Claude Code)
/plugin list

# Test with a simple task
/spec build a hello world CLI tool
/plan
/build
/test