| name | skill-creator |
| description | Create, edit, improve, or audit JaiClaw skills. Use when creating a new skill from scratch or when asked to improve, review, audit, or clean up an existing SKILL.md file. |
| alwaysInclude | false |
| requiredBins | [] |
| platforms | ["darwin","linux"] |
Skill Creator
Create and edit JaiClaw skills — modular knowledge packages that extend the agent with specialized workflows and domain expertise.
Skill Structure
skill-name/
└── SKILL.md (required)
SKILL.md Format
---
name: my-skill
description: What the skill does and when to use it
alwaysInclude: false
requiredBins: [tool-name]
platforms: [darwin, linux]
version: 1.0.0
tenantIds: []
---
# Skill Title
Instructions for the LLM...
Frontmatter Fields
| Field | Type | Default | Purpose |
|---|
name | string | required | Skill identifier |
description | string | required | One-line description (triggers skill selection) |
alwaysInclude | boolean | false | Include in every system prompt |
requiredBins | string[] | [] | Required CLI tools (checked via which) |
platforms | string[] | [] | Supported OSes: darwin, linux, windows |
version | string | "1.0.0" | Semantic version |
tenantIds | string[] | [] | Empty = all tenants; populated = restrict |
Design Principles
Concise is Key
The context window is shared. Only add context the LLM doesn't already have. Challenge each piece: "Does the agent really need this?"
Progressive Disclosure
- Metadata (name + description) — always in context
- SKILL.md body — loaded when skill triggers
- Keep body under 500 lines
Degrees of Freedom
- High freedom: text instructions when multiple approaches work
- Medium freedom: pseudocode when a preferred pattern exists
- Low freedom: exact scripts when consistency is critical
Creation Process
- Understand — gather concrete usage examples
- Plan — identify reusable resources (scripts, references)
- Create — write the SKILL.md with frontmatter and instructions
- Place — put in
src/main/resources/skills/{name}/SKILL.md (bundled) or .jaiclaw/skills/{name}/SKILL.md (workspace)
- Test — verify skill loads and triggers correctly
- Iterate — refine based on real usage
Naming
- Lowercase letters, digits, and hyphens only
- Short, verb-led phrases describing the action
- Namespace by tool when it improves clarity (e.g.,
gh-issues)
What NOT to Include
- README.md, CHANGELOG.md, or other auxiliary docs
- User-facing documentation about the creation process
- Information the LLM already knows well