OpenClaw Model Card
Use this skill to output OpenClaw model configuration in a consistent way.
Core workflow
- 1. Run text output:
python3 skills/openclaw-model-card/scripts/show-model-config.py --config /path/to/openclaw.json
- 2. If user wants a screenshot/image card, render image:
python3 skills/openclaw-model-card/scripts/show-model-config.py --config /path/to/openclaw.json --image ./model-card.png
- 3. If running in chat surfaces where terminal output is invisible to the user, send the script output via messaging tool instead of paraphrasing.
Rules
- - Do not handcraft model lists when this skill is requested; always use the script output.
- Prefer
--config explicitly for reproducibility. - Keep sensitive config files out of git; only publish generated, sanitized artifacts.
Notes
- -
md2img.js depends on wkhtmltoimage. - The script includes consistency checks (missing model refs, alias conflicts, suspicious contextWindow).
OpenClaw 模型卡片
使用此技能以一致的方式输出OpenClaw模型配置。
核心工作流程
- 1. 运行文本输出：
bash
python3 skills/openclaw-model-card/scripts/show-model-config.py --config /path/to/openclaw.json
- 2. 如果用户需要截图/图片卡片，渲染图像：
bash
python3 skills/openclaw-model-card/scripts/show-model-config.py --config /path/to/openclaw.json --image ./model-card.png
- 3. 如果在终端输出对用户不可见的聊天界面中运行，请通过消息工具发送脚本输出，而非转述。
规则
- - 当请求此技能时，不要手动构建模型列表；始终使用脚本输出。
- 优先明确使用--config以确保可复现性。
- 将敏感配置文件排除在git之外；仅发布经过处理的、已脱敏的产物。
备注
- - md2img.js依赖于wkhtmltoimage。
- 该脚本包含一致性检查（缺失的模型引用、别名冲突、可疑的contextWindow）。
