Set up MetaClaw to learn from your conversations.
Deploys a local conversation proxy that learns reusable instructions and schedules model training from live interactions.
When to use it
Use when setting up MetaClaw, configuring its operating modes or training backend, or deploying its proxy.
Give it a request to deploy MetaClaw; it guides interactive configuration and starts a local proxy for your client.
Reads the configured upstream LLM's API key from the environment.
Calendar-based scheduling requires Google Calendar OAuth consent and the Calendar API enabled in a Google Cloud project.
This skill
https://api.moonshot.cn/v1
Forwards conversations to the LLM
Tinker
Training through Tinker