Run persistent local agents for recurring work
Web page · https://agentusecases.ai/use-case/run-persistent-local-agents-for-recurring-work
The poster runs persistent agents locally to handle recurring tasks, saying a job that used to take three prompts now needs none.
First seen by Oriel on .
AI summary
The poster runs persistent local agents in isolated Linux containers for recurring work, reducing a task from three approval prompts to none while retaining safety checks and explicit completion criteria.
- Each worker gets its own Linux container, workspace, and sign-in, running locally through Codex on a Plus account.
- On every start, the poster verifies that containers cannot access the Mac or home network before allowing commands without repeated approvals.
- Research jobs prove completion by writing a file containing size, source count, hash, and confirmation that sources were listed rather than fact-checked.
- Agents stop when the laptop sleeps, plan limits remain the main capacity constraint, and the container boundary is strong but not guaranteed.
- The setup has only been tested with Colima on an Apple Silicon Mac.
#local agents #codex #linux containers #colima
Generated by gpt-5.6-sol. Check the original for details.
Aliases
| Source | External ID | URL |
|---|---|---|
| Agent Use Cases | https://agentusecases.ai/use-case/run-persistent-local-agents-for-recurring-work | https://agentusecases.ai/use-case/run-persistent-local-agents-for-recurring-work Read discussion |