What's the worst thing an AI agent left running on your machine?

I started building Clueprint after finding seven worktrees, four dev servers and a Docker stack I had no memory of starting. All spawned by Claude Code and Codex sessions from days earlier.

The app now watches all of it and tells me what's provably safe to stop. Launching here Sep 22.

Before then I'm curious: what's the worst thing an agent session ever left running for you? And is there anything you'd want a tool like this to catch that isn't obvious? Ports, containers and worktrees are covered.