Successfully set up OAuth - questions about scaling to multi-agent fleet
Just got OAuth working after a fresh install (thanks @Krill!). Now ready to scale from 1 agent to my 3-agent Docker fleet. Few questions:
1. Docker + OAuth setup:
I have CLU and Cortana running in separate Docker containers. What's the recommended approach?
Run clawdbot onboard inside each container and paste the same oat01 token?
Copy the auth config files from my working main agent?
Any Docker-specific gotchas?
2. Context overflow prevention:
All 3 agents will run 24/7. Best practices to avoid context overflow long-term?
ChromaDB integration?
Automated session resets?
Workspace file management strategies?
3. Model selection for dev agent:
CLU handles server setup, API integrations, and app development. Should he be:
Sonnet primary with Opus fallback (most dev work is straightforward)
Opus primary (complex architecture decisions)
Your experience?
4. Rate limit expectations & model recommendations:
Running on one Claude Pro OAuth ($20/month) with:
Jarvis: Coordinator (routes tasks, manages fleet)
CLU: Developer (server setup, APIs, builds)
Cortana: Content/marketing (social posts, copy)
What model configuration would you recommend for each to stay under rate limits while maintaining good performance?
Current architecture: All agents coordinate through custom dashboard, VPS-hosted.
1. Docker + OAuth setup:
I have CLU and Cortana running in separate Docker containers. What's the recommended approach?
Run clawdbot onboard inside each container and paste the same oat01 token?
Copy the auth config files from my working main agent?
Any Docker-specific gotchas?
2. Context overflow prevention:
All 3 agents will run 24/7. Best practices to avoid context overflow long-term?
ChromaDB integration?
Automated session resets?
Workspace file management strategies?
3. Model selection for dev agent:
CLU handles server setup, API integrations, and app development. Should he be:
Sonnet primary with Opus fallback (most dev work is straightforward)
Opus primary (complex architecture decisions)
Your experience?
4. Rate limit expectations & model recommendations:
Running on one Claude Pro OAuth ($20/month) with:
Jarvis: Coordinator (routes tasks, manages fleet)
CLU: Developer (server setup, APIs, builds)
Cortana: Content/marketing (social posts, copy)
What model configuration would you recommend for each to stay under rate limits while maintaining good performance?
Current architecture: All agents coordinate through custom dashboard, VPS-hosted.
The AI that actually does things. Emails, calendar, home automation - all from your favorite chat app. New shell, ready to help. The lobster way.
85,605Members
Resources
Similar Threads
Was this page helpful?
