Agent Skills
Agent Skills are modular capability systems that extend Claude’s functionality. Each Skill packages instructions, metadata, and optional resources (scripts, templates) that Claude uses automatically when relevant.
Why Use Skills
Skills are reusable, filesystem-based resources that provide Claude with domain-specific expertise: workflows, context, and best practices, transforming a general agent into a specialist.
Key Benefits:
- Specialize Claude: Customize capabilities for specific domain tasks
- Reduce Repetition: Create once, use automatically
- Compose Capabilities: Combine Skills to build complex workflows
Using Skills
Anthropic provides pre-built Agent Skills for common document tasks (PowerPoint, Excel, Word, PDF), and you can create your own custom Skills. Both work the same way—Claude uses them automatically when relevant.
Pre-built Agent Skills
Available to all claude.ai and Claude API users. See the Available Skills section for the complete list.
Custom Skills
Let you package domain expertise and organizational knowledge. Available across all Claude products: create in Claude Code, upload via API, or add in claude.ai settings.
How Skills Work
Skills leverage Claude’s VM environment to provide capabilities beyond pure prompting. Claude runs in a virtual machine with filesystem access, allowing Skills to exist as directories containing instructions, executable code, and reference materials.
Progressive Disclosure
Skills can contain three types of content, each loaded at different times:
Level 1: Metadata (Always Loaded)
- The Skill’s YAML frontmatter provides discovery information
- Claude loads this metadata at startup and includes it in the system prompt
- This lightweight approach means you can install many Skills without context penalty
Level 2: Instructions (Loaded When Triggered)
- The SKILL.md body contains procedural knowledge: workflows, best practices, and guidance
- When you request something matching the Skill description, Claude reads SKILL.md from the filesystem via bash
- Only then does the content enter the context window
Level 3: Resources and Code (Loaded On-Demand)
- Skills can bundle additional materials:
- Instructions: Additional markdown files
- Code: Executable scripts that Claude runs via bash
- Resources: Reference materials like database schemas, API docs, templates, or examples
| Level | Loading Time | Token Cost | Content |
|---|---|---|---|
| Level 1: Metadata | Always (at startup) | ~100 tokens/Skill | name and description in YAML frontmatter |
| Level 2: Instructions | When Skill triggers | <5k tokens | Instructions and guidance in SKILL.md body |
| Level 3+: Resources | On-demand | Virtually unlimited | Bundled files executed via bash, content not loaded to context |
Available Skills
Pre-built Agent Skills
The following pre-built Agent Skills are immediately available:
- PowerPoint (pptx): Create presentations, edit slides, analyze presentation content
- Excel (xlsx): Create spreadsheets, analyze data, generate reports with charts
- Word (docx): Create documents, edit content, format text
- PDF (pdf): Generate formatted PDF documents and reports
These Skills are available on Claude API and claude.ai. See the Quickstart tutorial to start using them in the API.
Official Skills Examples
Anthropic provides a rich collection of open-source Skills examples across multiple domains:
Document Processing:
- doc-coauthoring: Structured document collaboration workflow
- docx: Microsoft Word document processing (source available)
- pdf: PDF form filling and processing (source available)
- pptx: PowerPoint presentation creation (source available)
- xlsx: Excel spreadsheet processing (source available)
Creative & Design:
- frontend-design: Create unique production-ready frontend interfaces
- canvas-design: Canvas graphic design (includes 54 fonts)
- algorithmic-art: Generative algorithmic art
- theme-factory: 10 preset design themes
- slack-gif-creator: Slack GIF animation creation
Technical Development:
- mcp-builder: Model Context Protocol server building
- skill-creator: Agent Skills creation tool
- web-artifacts-builder: Web Artifacts builder
- webapp-testing: Web application automated testing
Enterprise & Communication:
- internal-comms: Enterprise internal communication templates
- brand-guidelines: Brand consistency guidance
For complete source code and detailed documentation, visit the Anthropic Skills repository.
Custom Skills Examples
If you want to create custom Skills, start with the local Quickstart and Best Practices pages in this site, then review the official Skills cookbook.
Skill Structure
Each Skill requires a SKILL.md file with YAML frontmatter:
---
name: Your Skill Name
description: Brief description of this Skill and when to use it
---
# Your Skill Name
## Instructions
[Clear step-by-step guidance for Claude to follow]
## Examples
[Concrete examples of using this Skill]Required Fields: name and description
Frontmatter Limits:
name: Maximum 64 charactersdescription: Maximum 1024 characters
Quick Navigation
Security Considerations
We strongly recommend using Skills only from trusted sources: ones you created yourself or obtained from Anthropic. Skills provide Claude with new capabilities through instructions and code—while this makes them powerful, it also means a malicious Skill could direct Claude to invoke tools or execute code in ways inconsistent with the Skill’s stated purpose.
Limitations and Constraints
Cross-Surface Availability
Custom Skills do not sync across surfaces. Skills uploaded to one surface are not automatically available on other surfaces.
Runtime Environment Constraints
Skills run in code execution containers with the following limitations:
- No Network Access: Skills cannot make external API calls or access the internet
- No Runtime Package Installation: Only pre-installed packages are available
- Only Pre-configured Dependencies: See Code Execution Tool documentation for the list of available packages