claude-skills-library-expert

# Claude Skills Library Expert > Skill by [ara.so](https://ara.so) — Claude Code Skills collection. Expert in using, installing, and contributing to the `alirezarezvani/claude-skills` repository — the most comprehensive open-source library of AI coding agent skills with 268+ production-ready skills across 9 domains (engineering, marketing, product, compliance, C-level advisory, etc.). ## What This Project Does The claude-skills library provides modular instruction packages (skills) that give AI coding agents domain expertise they don't have out of the box. Each skill includes: - **SKILL.md** — structured instructions, workflows, decision frameworks - **Python tools** — 373 CLI scripts (stdlib-only, zero pip installs) - **Reference docs** — templates, checklists, domain-specific knowledge **Multi-tool support:** Works natively with Claude Code, OpenAI Codex, Gemini CLI, and converts to 9+ other tools (Cursor, Aider, Windsurf, Kilo Code, OpenCode, Augment, Antigravity, Hermes Agent). ## Installation ### Claude Code (Recommended) ```bash # Add the marketplace /plugin marketplace add alirezarezvani/claude-skills # Install by domain /plugin install engineering-skills@claude-code-skills /plugin install marketing-skills@claude-code-skills /plugin install product-skills@claude-code-skills /plugin install c-level-skills@claude-code-skills # Install individual skills /plugin install skill-security-auditor@claude-code-skills /plugin install playwright-pro@claude-code-skills /plugin install self-improving-agent@claude-code-skills ``` ### Gemini CLI ```bash git clone https://github.com/alirezarezvani/claude-skills.git cd claude-skills ./scripts/gemini-install.sh # Activate a skill > activate_skill(name="senior-architect") ``` ### OpenAI Codex ```bash npx agent-skills-cli add alirezarezvani/claude-skills --agent codex ``` ### OpenClaw ```bash bash <(curl -s https://raw.githubusercontent.com/alirezarezvani/claude-skills/main/scripts/openclaw-install.sh) ``` ### Manual Installation ```bash git clone https://github.com/alirezarezvani/claude-skills.git # Copy skill folders to: # - Claude Code: ~/.claude/skills/ # - Codex: ~/.codex/skills/ # - Gemini CLI: ~/.gemini/skills/ ``` ## Converting Skills to Other Tools The library supports 12 AI coding tools via conversion scripts. ### Convert All Skills to All Tools ```bash cd claude-skills ./scripts/convert.sh --tool all ``` This generates tool-specific outputs in: - `.cursor/` — Cursor `.mdc` rules - `.aider/` — Aider `CONVENTIONS.md` - `.kilocode/` — Kilo Code rules - `.windsurf/` — Windsurf skills - `.opencode/` — OpenCode skills - `.augment/` — Augment rules - `~/.gemini/antigravity/` — Antigravity skills - `~/.hermes/` — Hermes Agent skills ### Install Converted Skills into a Project ```bash # Install with confirmation ./scripts/install.sh --tool cursor --target /path/to/project # Force install without confirmation ./scripts/install.sh --tool aider --target . --force ``` ### Tool-Specific Conversion ```bash # Cursor ./scripts/convert.sh --tool cursor ./scripts/install.sh --tool cursor --target . # Aider ./scripts/convert.sh --tool aider ./scripts/install.sh --tool aider --target . # Windsurf ./scripts/convert.sh --tool windsurf ./scripts/install.sh --tool windsurf --target . # Hermes Agent python scripts/sync-hermes-skills.py --verbose ``` ### Verify Installation ```bash # Cursor find .cursor/rules -name "*.mdc" | wc -l # Should show 268 # Aider cat CONVENTIONS.md # Should contain all skills # Windsurf ls -1 .windsurf/skills/ | wc -l # Should show 268 ``` ## Key Skills and Domains ### Engineering (32 core + 40 POWERFUL) ```bash # Core skills /plugin install senior-architect@claude-code-skills /plugin install senior-frontend@claude-code-skills /plugin install senior-backend@claude-code-skills /plugin install devops-engineer@claude-code-skills /plugin install playwright-pro@claude-code-skills # POWERFUL tier /plugin install agent-designer@claude-code-skills /plugin install rag-architect@claude-code-skills /plugin install database-designer@claude-code-skills /plugin install ci-cd-pipeline-builder@claude-code-skills /plugin install skill-security-auditor@claude-code-skills ``` ### Marketing (44 skills) ```bash /plugin install marketing-skills@claude-code-skills # Individual pods # Content (8), SEO (5), CRO (6), Channels (6), Growth (4), Intelligence (4), Sales (2) ``` ### Product (13 skills) ```bash /plugin install product-skills@claude-code-skills # Includes: product-manager, agile-po, ux-researcher, ui-design, analytics-tracking, etc. ``` ### C-Level Advisory (28 skills) ```bash /plugin install c-level-skills@claude-code-skills # Full C-suite: CTO, CMO, CFO, CRO, CPO, COO, CHRO, CISO, GC, CAIO ``` ## Using Personas Personas are pre-configured agent identities with curated skill loadouts and distinct communication styles. ### Available Personas - **startup-cto** — Architecture decisions, tech stack, team building - **growth-marketer** — Content-led growth, launch strategy, channel optimization - **solo-founder** — Cross-domain, MVP building, wearing all hats ### Install a Persona ```bash # Claude Code cp agents/personas/startup-cto.md ~/.claude/agents/ # For other tools, convert first ./scripts/convert.sh --tool cursor ``` ### Use a Persona ```markdown # In your conversation with the AI agent: "I want you to act as the Startup CTO persona. Review my architecture decisions for this new feature." ``` ## Orchestration Patterns Orchestration coordinates personas, skills, and agents across domain boundaries. ### Four Patterns **1. Solo Sprint** — Switch personas across project phases ```markdown Week 1-2: startup-cto + aws-solution-architect Week 3-4: growth-marketer + launch-strategy Week 5-6: solo-founder + analytics-tracking ``` **2. Domain Deep-Dive** — One persona + multiple stacked skills ```bash # Example: Security audit /activate persona startup-cto /use skill-security-auditor /use aws-solution-architect /use compliance-automation ``` **3. Multi-Agent Handoff** — Personas review each other's output ```markdown 1. startup-cto designs architecture 2. devops-engineer reviews deployment strategy 3. skill-security-auditor audits the plan ``` **4. Skill Chain** — Sequential skills, no persona needed ```bash # Content pipeline /use copywriting → /use seo-content-optimizer → /use email-sequence ``` See `orchestration/ORCHESTRATION.md` for full protocol and examples. ## Creating a New Skill ### Skill Structure ``` my-skill/ ├── SKILL.md # Main instruction file ├── scripts/ # Python tools (optional) │ └── tool.py ├── references/ # Documentation (optional) │ └── guide.md └── templates/ # Templates (optional) └── template.txt ``` ### SKILL.md Template ```yaml --- name: my-skill description: One-line description of what this skill does triggers: - "phrase a user might say" - "another trigger phrase" - "how do I..." - "show me..." --- # My Skill > Skill by [ara.so](https://ara.so) — Claude Code Skills collection. ## Overview What this skill does and when to use it. ## Key Capabilities - Capability 1 - Capability 2 ## Usage How to use this skill with code examples. ## Workflows Step-by-step workflows. ## Troubleshooting Common issues and solutions. ``` ### Python Tool Template ```python #!/usr/bin/env python3 """ Description of what this tool does. """ import sys import json import os def main(): # Use environment variables for configuration config = os.getenv('MY_CONFIG', 'default_value') # Your tool logic here result = {"status": "success"} # Output JSON for easy parsing print(json.dumps(result, indent=2)) if __name__ == "__main__": main() ``` ### Add Your Skill ```bash # 1. Create skill directory mkdir -p my-skill/scripts # 2. Write SKILL.md and scripts/tool.py # 3. Test locally cp -r my-skill ~/.claude/skills/ # 4. (Optional) Submit PR to claude-skills repo git checkout -b add-my-skill git add my-skill/ git commit -m "Add my-skill" git push origin add-my-skill ``` ## Common Patterns ### Using Python Tools from Skills All 373 Python tools are stdlib-only and can run anywhere: ```python #!/usr/bin/env python3 # Example: engineering-team/playwright-pro/scripts/generate-test.py import sys import json def generate_test(page_url, selectors): """Generate a Playwright test from URL and selectors.""" test_code = f""" import {{ test, expect }} from '@playwright/test'; test('test {page_url}', async ({{ page }}) => {{ await page.goto('{page_url}'); await expect(page.locator('{selectors[0]}')).toBeVisible(); }}); """ return test_code if __name__ == "__main__": url = sys.argv[1] if len(sys.argv) > 1 else "https://example.com" selectors = sys.argv[2:] if len(sys.argv) > 2 else ["body"] print(generate_test(url, selectors)) ``` **Usage:** ```bash python engineering-team/playwright-pro/scripts/generate-test.py \ "https://myapp.com/login" \ "input[name='email']" \ "button[type='submit']" ``` ### Stacking Skills ```bash # Claude Code /use senior-architect /use aws-solution-architect /use skill-security-auditor # Then ask: "Review my architecture for a multi-tenant SaaS with SOC 2 compliance requirements." ``` ### Using Skills with Code Generation ```markdown User: "Generate a REST API with CRUD operations for a blog post resource." AI (with senior-backend skill): - Analyzes requirements - Follows RESTful patterns from skill - Generates code with proper validation - Includes security best practices - Adds tests and documentation ``` ## Troubleshooting ### Skills Not Loading **Claude Code:** ```bash # Check installation ls ~/.claude/skills/ # Reinstall /plugin uninstall my-skill@claude-code-skills /plugin install my-skill@claude-code-skills ``` **Gemini CLI:** ```bash # Check installation ls ~/.gemini/skills/ # Reinstall ./scripts/gemini-install.sh ``` ### Conversion Script Fails ```bash # Ensure you're in the repo root cd /path/to/claude-skills # Check Python version (requires 3.7+) python3 --version # Run with verbose output ./scripts/convert.sh --tool cursor --verbose ``` ### Python Tools Not Executable ```bash # Make scripts executable find . -name "*.py" -type f -exec chmod +x {} \; # Or run with python3 explicitly python3 engineering-team/playwright-pro/scripts/generate-test.py ``` ### Persona Not Activating ```bash # Ensure persona file is in correct location # Claude Code cp agents/personas/startup-cto.md ~/.claude/agents/ # For other tools, convert first ./scripts/convert.sh --tool cursor # Verify file exists cat ~/.claude/agents/startup-cto.md ``` ### Skill Security Concerns Use the `skill-security-auditor` before installing third-party skills: ```bash # Install the security auditor /plugin install skill-security-auditor@claude-code-skills # Audit a skill directory /use skill-security-auditor "Audit the skills in /path/to/untrusted-skills/" ``` ## Configuration ### Environment Variables Skills respect these environment variables: ```bash # General export CLAUDE_SKILLS_PATH="$HOME/.claude/skills" export GEMINI_SKILLS_PATH="$HOME/.gemini/skills" # For tools with API integration export OPENAI_API_KEY="your-key-here" export ANTHROPIC_API_KEY="your-key-here" # Project-specific export PROJECT_ROOT="/path/to/project" export DOCS_OUTPUT_DIR="./docs" ``` ### Tool-Specific Configuration