Create Skill

1---2name: create-skill3description: Guide for creating effective skills following best practices. Use when creating or updating skills that extend agent capabilities.4---5 6# Create Skill7 8Guide for creating effective skills that extend agent capabilities with specialized knowledge, workflows, and tool integrations.9 10## About Skills11 12Skills are modular, self-contained packages that extend agent capabilities by providing specialized knowledge, workflows, and tools. Think of them as "onboarding guides" for specific domains or tasks.13 14### What Skills Provide15 161. Specialized workflows - Multi-step procedures for specific domains172. Tool integrations - Instructions for working with specific file formats or APIs183. Domain expertise - Company-specific knowledge, schemas, business logic194. Bundled resources - Scripts, references, and assets for complex and repetitive tasks20 21## Progressive Disclosure Principle22 23**The 200-line rule is critical.** SKILL.md must be under 200 lines. If you need more, split content into `references/` files.24 25### Three-Level Loading System26 271. **Metadata (name + description)** - Always in context (~100 words)282. **SKILL.md body** - When skill triggers (<200 lines, ideally <500 lines for optimal performance)293. **Bundled resources** - As needed by agent (unlimited)30 31### Why Progressive Disclosure Matters32 33- 85% reduction in initial context load34- Activation times drop from 500ms+ to under 100ms35- Agent loads only what's needed, when it's needed36- Skills remain maintainable and focused37 38## Skill Structure39 40```41skill-name/42├── SKILL.md (required, <200 lines)43│   ├── YAML frontmatter metadata (required)44│   │   ├── name: (required)45│   │   └── description: (required)46│   └── Markdown instructions (required)47└── Bundled Resources (optional)48    ├── scripts/          - Executable code49    ├── references/       - Documentation loaded as needed50    └── assets/           - Files used in output51```52 53## Core Principles54 55### Concise is Key56 57The context window is a shared resource. Your skill shares it with everything else the agent needs. Be concise and challenge each piece of information:58- Does the agent really need this explanation?59- Can I assume the agent knows this?60- Does this paragraph justify its token cost?61 62### Set Appropriate Degrees of Freedom63 64- **High freedom**: Text-based instructions for multiple valid approaches65- **Medium freedom**: Pseudocode or scripts with parameters66- **Low freedom**: Specific scripts with few/no parameters for fragile operations67 68### Test with All Models69 70Skills act as additions to models, so effectiveness depends on the underlying model. Test your skill with all models you plan to use it with.71 72## References73 74For detailed guidance, see:75- `references/progressive-disclosure.md` - 200-line rule and references pattern76- `references/skill-structure.md` - SKILL.md format and frontmatter details77- `references/examples.md` - Good skill examples78- `references/best-practices.md` - Comprehensive best practices guide