Skills are reusable instruction documents that give agents specialized knowledge or guidelines.Key properties:
Property
Description
namespace / name
Unique identifier
description
What the skill helps with (shown to the LLM as the tool description)
content
Markdown instructions
Two modes:
Mode
Behavior
Best for
Preloaded (preload: true)
Injected into the system prompt
Tone guidelines, safety rules, persona traits
Progressive (preload: false)
Exposed as a tool the LLM calls when needed
Research methods, domain expertise, task-specific instructions
Example agent configuration:
enabled_skills: - skill: "default/tone_guidelines" preload: true # Always present in system prompt - skill: "default/web_research" preload: false # LLM decides when to retrieve it
Endpoints:
POST /api/v1/skills # Create skillGET /api/v1/skills # List skillsGET /api/v1/skills/{namespace}/{name} # Get skillPUT /api/v1/skills/{namespace}/{name} # Update skillDELETE /api/v1/skills/{namespace}/{name} # Delete skill
