xagent-integration
Run Claude Code inside Muse's computer through a polling bridge
At a glance
A developer connected a browser terminal to Claude Code running on a Muse agent's Linux VM. This workflow is listed under agent-integration, terminal, session-management, scheduled-tasks.
- Agents / tools
- Meta Muse, Claude Code, xterm.js
Building a tool for these workflows? See weekly rates.
What they did
With no inbound ports available, the developer built an xterm.js interface and used a scheduled Muse task to check for prompts every five seconds, invoke claude -p and return the output. A stored session ID per thread enabled follow-up requests with --resume.
How it went
The author reports interacting with Claude Code through the bridge and describes the backend as about fifteen lines of glue.
Worth knowing
The five-second polling loop was the author's workaround for the VM's lack of inbound ports.
Save ideas in this browser. Upvotes count across the board.