Terminal Skills

An open-source library of AI agent skills following the Agent Skills open standard. Skills work across Claude Code, OpenAI Codex, Gemini CLI, Cursor, and other AI-powered development tools.

Browse the full catalog at terminalskills.io.

Each skill is a standalone SKILL.md file. Install skills into your project using one of the following methods.

npx terminal-skills install pdf-analyzer

Or using curl:

curl -sL https://raw.githubusercontent.com/terminal-skills/skills/main/skills/pdf-analyzer/SKILL.md -o .claude/skills/pdf-analyzer.md

curl -sL https://raw.githubusercontent.com/terminal-skills/skills/main/skills/pdf-analyzer/SKILL.md -o .codex/skills/pdf-analyzer.md

curl -sL https://raw.githubusercontent.com/terminal-skills/skills/main/skills/pdf-analyzer/SKILL.md -o .gemini/skills/pdf-analyzer.md

curl -sL https://raw.githubusercontent.com/terminal-skills/skills/main/skills/pdf-analyzer/SKILL.md -o .cursor/skills/pdf-analyzer.md

Replace pdf-analyzer with the skill name you want to install.

Two of these skills are backed by a hosted MCP server — so your agent can call them as live tools, not just read instructions:

There are two ways to connect either server — no plugin install required:

1. As a connector (claude.ai web/desktop — OAuth, one click, no key to copy)

In Claude: Settings → Connectors → Add custom connector, paste the server URL, and sign in with your account. Claude discovers the tools and handles auth for you.

2. As an MCP server in a CLI (Claude Code, Codex, Cursor, …) with a token

Create a token on your account page, then add the server to your MCP config:

{
  "mcpServers": {
    "terminal-skills": {
      "type": "http",
      "url": "https://api.terminalskills.io/api/mcp",
      "headers": { "Authorization": "Bearer YOUR_TOKEN" }
    }
  }
}

For SEOG, use https://api.seog.ai/mcp and a token from seog.ai. The underlying skill (seog) still installs on its own with npx terminal-skills install seog if you only want the instructions.

Step-by-step guides for common workflows:

• Analyze PDF Documents
• Process Excel Data
• Automate Code Reviews
• Write Better Commits
• Package Repeated AI Agent Workflows into Skills
• Test API Endpoints
• Manage Docker Containers
• Scrape Web Data
• Create Data Visualizations
• Generate Documentation
• Optimize SQL Queries
• Test AI Prompts Systematically
• Route AI Model Requests Through RouterBase
• Manage Monorepo Dependencies
• Validate ETL Data Quality
• Automate Release Notes
• Debug Production Incidents
• Automate Local SEO Monitoring for Client Businesses

See CONTRIBUTING.md for guidelines on creating and submitting new skills.

Apache-2.0. See LICENSE.