Agent Skill
2/7/2026skill-creator
Guide for creating effective skills. Use when users want to create a new skill that extends Claude's capabilities with specialized knowledge, workflows, or tool integrations.
B
baltsat
0GitHub Stars
1Views
npx skills add Baltsat/box
SKILL.md
| Name | skill-creator |
| Description | Guide for creating effective skills. Use when users want to create a new skill that extends Claude's capabilities with specialized knowledge, workflows, or tool integrations. |
name: skill-creator description: Guide for creating effective skills. Use when users want to create a new skill that extends Claude's capabilities with specialized knowledge, workflows, or tool integrations.
Skill Creator
Guide for creating effective Claude Code skills.
What Skills Provide
- Specialized workflows - Multi-step procedures for specific domains
- Tool integrations - Instructions for working with specific file formats or APIs
- Domain expertise - Company-specific knowledge, schemas, business logic
- Bundled resources - Scripts, references, and assets for complex tasks
Anatomy of a Skill
skill-name/
├── SKILL.md (required)
│ ├── YAML frontmatter (name, description)
│ └── Markdown instructions
└── Optional Resources
├── scripts/ - Executable code
├── references/ - Documentation
└── assets/ - Templates, icons
SKILL.md Structure
Frontmatter (required)
---
name: my-skill
description: What this skill does and when to use it. Include trigger phrases.
---
Important: The description is the primary triggering mechanism. Include both what it does AND when to use it.
Body
Instructions and guidance for using the skill. Only loaded AFTER the skill triggers.
Core Principles
Concise is Key
Context window is shared. Only add what Claude doesn't already know. Challenge each piece: "Does Claude really need this?"
Prefer concise examples over verbose explanations.
Degrees of Freedom
Match specificity to task fragility:
- High freedom (text instructions): Multiple approaches valid
- Medium freedom (pseudocode): Preferred pattern exists
- Low freedom (specific scripts): Fragile/error-prone operations
Progressive Disclosure
Three-level loading system:
- Metadata (name + description) - Always in context (~100 words)
- SKILL.md body - When skill triggers (<5k words)
- Bundled resources - As needed
What NOT to Include
- README.md, INSTALLATION_GUIDE.md, CHANGELOG.md
- User-facing documentation
- Setup/testing procedures
- Content Claude can infer from codebase
Creation Process
- Understand - Get concrete examples of how skill will be used
- Plan - Identify reusable resources (scripts, references, assets)
- Initialize - Create skill directory with SKILL.md
- Edit - Implement resources and write instructions
- Test - Use the skill on real tasks
- Iterate - Improve based on usage
Validation
Skills must have SKILL.md with frontmatter containing name and description:
# Validate skill
./sync.sh validate
Installing Skills
# Add to ~/.claude/skills/
ln -sfn /path/to/skill ~/.claude/skills/skill-name
Or use npx skills CLI:
npx skills add owner/repo@skill-name
Skills Info
Original Name:skill-creatorAuthor:baltsat
Download