LIBERO Fix Loop
Use this skill for Experiment 1. Start with INSTRUCTIONS.md.
Run Order
- Follow main-agent-prompt.md as coordinator.
- Fill subagent-prompt.md once per task.
- Each worker follows skills/task-exploration.md, generates its own initial code, then uses the original failure-by-failure debug loop.
- Use clean-task-slate.md before reruns.
- Promote Stage 1-supported patterns into ../skills/ and record each update with
scripts/libero/record_skill_promotion.pybefore dispatching the next Stage 1 task. Held-out outcomes never drive skill edits.
No external baseline code or baseline output directory is used.