Create modular, self-contained skills that extend agent capabilities.
1. Understand → 2. Plan → 3. Init → 4. Edit → 5. Package → 6. Iterate
Ask clarifying questions if needed. Understand:
When to skip: When usage patterns are already clear.
Identify what to bundle:
Always use init_skill.py — never create manually:
python3 <skill-creator>/scripts/init_skill.py <name> --path <output-dir> --resources scripts,references,assets
Example:
python3 /usr/local/lib/node_modules/openclaw/skills/skill-creator/scripts/init_skill.py my-skill --path /var/root/.openclaw/workspace/skills --resources scripts,references
---
name: skill-name
description: Clear description of what it does AND when to trigger it. Include Chinese triggers too.
---
# Skill Name
## Quick Start
## Details...
pdf-edit, list-skills)references/python3 <skill-creator>/scripts/package_skill.py <path/to/skill-folder>
Output: .skill file (zip format).
Validation runs automatically before packaging.
Test with real tasks, gather feedback, improve.
skill-name/
├── SKILL.md # Required: name + description + instructions
├── scripts/ # Optional: executable code
├── references/ # Optional: load-as-needed docs
└── assets/ # Optional: output files
| Resource | When to Use |
|---|---|
| ---------- | ------------- |
scripts/ | Same code rewritten repeatedly |
references/ | Large docs the model should read on demand |
assets/ | Templates, images, boilerplate |
Keep SKILL.md under 500 lines. Move details to references/.
共 1 个版本