Agent Skills are a simple, open format for giving agents new capabilities and expertise.
Skills are folders of instructions, scripts, and resources that agents can discover and use to perform better at specific tasks. Write once, use everywhere.
pip install skills-refOr using uv:
uv add skills-ref# Validate a skill
agentskills validate path/to/skill
# Read skill properties (outputs JSON)
agentskills read-properties path/to/skill
# Generate <available_skills> XML for agent prompts
agentskills to-prompt path/to/skill-a path/to/skill-bfrom pathlib import Path
from skills_ref import validate, read_properties, to_prompt
# Validate a skill directory
errors = validate(Path("my-skill"))
if errors:
print("Validation errors:", errors)
else:
print("Valid skill!")
# Read skill properties
props = read_properties(Path("my-skill"))
print(f"Skill: {props.name}")
print(f"Description: {props.description}")
# Generate prompt XML for agent integration
prompt = to_prompt([Path("skill-a"), Path("skill-b")])
print(prompt)- Validation: Check skills for proper format and required fields
- Parsing: Extract skill metadata from SKILL.md frontmatter
- Prompt Generation: Create XML blocks for agent system prompts
- i18n Support: Full Unicode support for international skill names
- CLI Tools: Command-line utilities for skill management
- Full Documentation - Guides and tutorials
- Specification - Format details
- Example Skills - See what's possible
Agent Skills is an open format maintained by Anthropic and open to contributions from the community.
Apache 2.0