Curious how others are handling isolation when running coding agents like Claude Code or Codex. Are you running them directly on your machine? Git worktrees? Docker? Something else?
I built SuperHQ because I wanted real VM isolation for each agent session. But I am curious what setups others are using and what pain points you are running into.
SuperHQ orchestrates Claude Code, Codex, and Pi coding agents inside isolated microVMs, with a secure auth gateway that keeps your API keys out of the sandbox.
