Api Model list updater

该 Skill 读取 OpenClaw 配置中的提供方列表，向所选提供方查询可用模型，由用户挑选后把模型定义写入 models.providers，并在 agents.defaults.models 中生成别名条目，写入前先做配置备份。附带一个 Python 脚本，但文档未说明其实现细节。

da-kaine

热度 25.14

最近更新 2026-05-11 17:30:03

分类：自动化依赖：代码执行 / 读写文件 / 第三方 API应用场景：工作流 / API 集成 / 自动化脚本

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add model-catalog-updater

Metadata

name: model-catalog-updater
version: 1.2.0
description: Query available models from your configured providers and add them to OpenClaw config
homepage: https://github.com/openclaw/skills/model-catalog
metadata: {"model-catalog-updater":{"emoji":"📊","category":"developer-tools","config_path":"C:\\Users\\zebuM\\.openclaw\\openclaw.json","backup_name":"openclawworking.json","defaults":{"contextWindow":128000,"maxTokens":8192,"reasoningModelPatterns":["deepseek-r1","qwq","o1","o3","reasoning","think"]}}}
commands: [{"name":"model-catalog-updater","description":"Fetch and add models from a provider to your config","options":[{"name":"provider","description":"Select a provider to fetch models from","type":"string","required":true,"choices":[{"name":"qwen-portal","value":"qwen-portal"},{"name":"lmstudio","value":"lmstudio"},{"name":"minimax-portal","value":"minimax-portal"},{"name":"xai","value":"xai"},{"name":"openrouter","value":"openrouter"},{"name":"modal-direct","value":"modal-direct"},{"name":"minimax","value":"minimax"},{"name":"google","value":"google"},{"name":"openai-codex","value":"openai-codex"},{"name":"github-copilot","value":"github-copilot"},{"name":"groq","value":"groq"},{"name":"opencode","value":"opencode"},{"name":"zai","value":"zai"}]}]}]

Model Catalog Updater

Fetch available models from your configured API providers and add them to your OpenClaw config.

How It Works

When invoked, this skill:

Reads your config → Loads providers from openclaw.json
Presents choices → Shows numbered list of your configured providers
User selects provider → Queries /v1/models from that provider
Shows available models → Lists models with IDs
User picks models → Select which to add (comma-separated or 'all')
Creates backup → Saves openclawworking.json before any edits
Updates config → Adds models to both:
- models.providers.{provider}.models[] — model definitions
- agents.defaults.models.{provider/model-id} — alias entries

Usage

Slash command:

/model-catalog-updater provider:<dropdown>

Natural language:

"show me available models"
"add models from openrouter"
"what models are on lmstudio"

Configuration

Setting	Value
Config path	`C:\Users\zebuM\.openclaw\openclaw.json`
Backup file	`openclawworking.json` (same directory)
Default context window	128000
Default max tokens	8192

Reasoning Detection

Models are auto-detected as reasoning models if their ID contains:

deepseek-r1
qwq
o1
o3
reasoning
think

Model Entry Format

Added to models.providers.{provider}.models:

{
  "id": "model-id",
  "name": "Display Name",
  "reasoning": false,
  "input": ["text"],
  "cost": { "input": 0, "output": 0, "cacheRead": 0, "cacheWrite": 0 },
  "contextWindow": 128000,
  "maxTokens": 8192
}

Alias Entry Format

Added to agents.defaults.models:

"provider/model-id": {}

(Empty object — user can add "alias": "short-name" manually if desired)

Example Session

📊 Model Catalog

Your configured providers:
1. qwen-portal (https://portal.qwen.ai/v1)
2. lmstudio (http://192.168.56.1:1234/v1)
3. xai (https://api.x.ai/v1)
4. openrouter (https://openrouter.ai/api/v1)
5. modal-direct (https://api.us-west-2.modal.direct/v1)
6. cloudflare-ai-gateway
7. minimax

Select provider [1-7]: 3

Fetching models from xai...

Available models:
1. grok-4
2. grok-4-vision
3. grok-2-1212

Select models to add (comma-separated, or 'all'): 1,2

✅ Backup created: openclawworking.json
✅ Added to models.providers.xai.models:
   - grok-4
   - grok-4-vision
✅ Added to agents.defaults.models:
   - xai/grok-4
   - xai/grok-4-vision

Notes

Some providers don't expose /v1/models (e.g., Cloudflare AI Gateway)
Vision models get "input": ["text", "image"] automatically if ID contains "vision" or "multimodal"
Restart OpenClaw after adding models to apply changes