This article covers skills on Claude. Recently, you can also use skills with chatGPT. We do currrently not include documentation of chatGPT.
Skills are folders containing instructions, scripts, and resources that Claude loads dynamically to improve performance on specialized tasks. They function as "onboarding materials" that package domain expertise, making Claude a specialist for specific workflows without requiring you to re-explain processes each time.
Claude automatically detects when a skill is relevant based on your request—no explicit invocation needed. Skills work across Claude apps (web/desktop), Claude Code, and the API.
Consistency and repeatability: Once defined, a skill ensures Claude follows the same process, output format, and quality standards every time.
Context efficiency: Skills use progressive disclosure—Claude loads only the metadata (name/description) initially, then accesses full instructions only when needed. This means you can have many skills installed without context overhead.
Composable: Multiple skills can work together. Claude identifies which ones are needed and coordinates their use automatically.
Portable: The same skill format works across Claude apps, Claude Code, and the API.
Executable code: Skills can include Python or bash scripts for tasks where programmatic execution is more reliable than token generation (e.g., document generation with specific formatting).
Prerequisites: You need a Pro, Max, Team, or Enterprise plan, and "Code execution and file creation" must be enabled.
Steps:
Skill structure (minimum):
skill-name/
├── SKILL.md (required)
The SKILL.md file requires YAML frontmatter with name and description fields. The description is critical—it determines when Claude invokes the skill.
Security note: Only install skills from trusted sources. Skills can execute code and install packages, so review third-party skills before enabling them.