Model Healthcheck
by @xmanrui
Test all configured models for availability. Activate when user says "test models", "check models", "model healthcheck", "测试模型", "测试所有模型", "检查模型", or "模型健康检查".
clawhub install model-healthcheck📖 About This Skill
name: model-healthcheck description: Test all configured models for availability. Activate when user says "test models", "check models", "model healthcheck", "测试模型", "测试所有模型", "检查模型", or "模型健康检查".
Model Healthcheck Skill
When the user requests a model test, follow these steps:
Steps
1. Get all model list
Use gateway config.get to read the current config and extract all provider/model combinations from models.providers.
2. Test all models concurrently
For each model, use sessions_spawn to run a test:
- task: "Reply with exactly: model ok. Only reply these two words."
- model: the corresponding provider/model-id
- runTimeoutSeconds: 30
- cleanup: "delete"
Fire all tests concurrently (multiple sessions_spawn calls at once), do not test sequentially.
3. Collect results and summarize Wait for all tests to complete, then summarize as a list:
- provider/model-id — ✅ OK / ❌ Failed (Xs) — error message if any
Fields: - Status: normal reply = ✅, timeout or error = ❌ - Duration: time from spawn to result - Error: if failed, show the specific error (API error, timeout, auth failure, etc.)
4. Output results Send the summary to the user, with a final tally: - Total models tested: X - Passed: X - Failed: X (list failed model names and reasons)