Enable fully autonomous agents
Question
Hello there, thank you for your hard work. I wanted to ask if is it possible to enable fully autonomous agents given a large tasklist in a .md file. I tried but it implements until an arbitrary step and stops. Thank you in advance
This issue might be a duplicate of existing issues. Please check:
- #6792: Task Tool Timeouts & Early Termination in Multi-Agent Conductor Pattern
Feel free to ignore if none of these address your specific case.
unfortunately, it does not work. I made a .md with tasks to do but model always stops after a little while asking if it can continue.
I don't think this is an OpenCode issue. In my experience, this issue seems to be model dependent. I often have GPT OSS 120b stop and ask to continue, while GLM 4.5 Air will work for hours on end until the task is complete.
Really? Are you using a particular prompt? I tried claude sonnet 4.5 and minimax-m2.1 free with spec-kit for an experiment and it always returned asking to continue after a little while
Depends on model, not an opencode issue, also consider doing ralph methods