Best way to give OpenClaw agents access to Claude Code / verified skills?
Hey @Krill 🦐 - I'm running 3 OpenClaw agents on a VPS and want to give them access to Claude Code's capabilities (verified skills, MCP servers, better code generation).
Current approach:
• Installed Claude Code (2.1.71) on the VPS
• Planning to OAuth authenticate manually
• Agents would call it via exec: "claude 'do something'"
My questions:
1. Is Claude Code designed for programmatic agent calls, or is it purely for interactive human dev use?
2. Will OAuth tokens persist across multiple agent exec calls?
3. Is there a better way to give agents these capabilities (direct MCP integration, API calls, custom skills)?
Context:
• 3 agents (Jarvis, CLU, Cortana) running under same Linux user
• Want coding superpowers (verified skills, MCP access) for CLU specifically
• Open to any approach that actually works in production
What's the recommended pattern here? Thanks!
Current approach:
• Installed Claude Code (2.1.71) on the VPS
• Planning to OAuth authenticate manually
• Agents would call it via exec: "claude 'do something'"
My questions:
1. Is Claude Code designed for programmatic agent calls, or is it purely for interactive human dev use?
2. Will OAuth tokens persist across multiple agent exec calls?
3. Is there a better way to give agents these capabilities (direct MCP integration, API calls, custom skills)?
Context:
• 3 agents (Jarvis, CLU, Cortana) running under same Linux user
• Want coding superpowers (verified skills, MCP access) for CLU specifically
• Open to any approach that actually works in production
What's the recommended pattern here? Thanks!
The AI that actually does things. Emails, calendar, home automation - all from your favorite chat app. New shell, ready to help. The lobster way.
130,529Members
Resources
Similar Threads
Was this page helpful?
