# Openclaw Model Card

> Generate OpenClaw model inventory and model-card images from openclaw.json. Use when asked to list all configured models, verify default/fallback chains, or render/share a visual model card screenshot.

- Skill: `dvcrn/openclaw-model-card` (Agent Skill, multi-file: 3 files)
- Install (CLI): `npx skillmds@latest add dvcrn/openclaw-model-card`
- Raw SKILL.md: https://api.skillmd.com/api/skills/dvcrn/openclaw-model-card/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: dvcrn (https://skillmd.com/u/dvcrn)
- Updated: 2026-09-08
- Page: https://skillmd.com/skills/dvcrn/openclaw-model-card

---


# OpenClaw Model Card

Use this skill to output OpenClaw model configuration in a consistent way.

## Core workflow

1. Run text output:
   ```bash
   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:
   ```bash
   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).