Skill Creator Tool (Free Edition) for Openclaw

1. Verify your runtime environment.

python3 --version

2. Install the minimal Python dependency used in the example workflow.

pip install requests

3. Prepare a working directory for the skill content.

mkdir -p my-skill
cd my-skill

4. Create or edit SKILL.md using the markdown structure expected by the tool. The free edition is markdown-native and does not require an extra API key.

5. Run the skill creation or validation workflow from your AI agent environment. Openclaw Skills supports agent platforms that can read SKILL.md and execute commands, including Claude Code, Cursor, Codex, and Gemini CLI.

6. Confirm the output format and execution log. The default output is structured JSON, and the free edition is intended for single-task, personal usage.

The tool organizes data around a markdown-first skill package plus structured execution metadata.

Core input fields

SKILL.md taxonomy

• slug: machine-friendly identifier for the skill
• name: canonical skill name
• version: semantic version string, such as 1.0.0
• displayName: user-facing display title
• summary: short functional description
• license: licensing metadata
• edition: release tier, such as free
• description: expanded narrative including keywords and use scenarios
• tags: capability and discovery keywords
• tools: supported execution tools such as read, write, and exec
• homepage: product or platform homepage
• pricing_tier, pricing_model, suggested_price: commercial metadata for packaging

Runtime output schema

{
  "success": true,
  "data": {
    "result": "...",
    "execution_time": "...",
    "metadata": {
      "version": "1.0",
      "processor": "skill creator"
    }
  },
  "execution_log": ["..."],
  "error": null
}

Operational format notes

• Input/output is standardized for create, query, and export workflows.
• The tool favors JSON output for downstream automation.
• The free edition is optimized for single-file processing and basic verification.
• SKILL.md is the primary artifact; generated content centers on markdown structure and metadata correctness.