Compile Once, Run Cheap: Frozen, Portable Skills for Embodied Agents, and What Recursive Self-Improvement Adds

Key points

• We study a third route: a strong general model acts only as an offline author that writes, runs, diagnoses and repairs a manipulation skill in simulation until it passes an exam on layouts drawn after freezing; the frozen skill then runs as code alone or under a cheap caller that reads only its description.1
• We call this skill-level recursive self-improvement Skill-RSI and specify a compile protocol (isolated, audited author; hidden validation layouts; no-feedback and inherit-only controls; tamper-evident ledger) and a reuse protocol that runs the identical files with executors from another model family.1
• On 7 RoboTwin 2.0 tasks with cluttered tables, Claude Code compiled 4 tasks into 7 frozen skills, 5 from a first draft needing only qualification, at USD 4.29-8.22 of author cost each (USD 55 per skill counting the 13 failed campaigns).1
• On hidden layouts, repair that inherits the previous version beat from-scratch rewriting in 8 of 8 controlled campaigns (sign test p = 0.004); an inherit-only lineage attributes most of the gap to inheritance, with development feedback adding a little (3 of 3, p = 0.125).1

Sources (1)

• [1]Compile Once, Run Cheap: Frozen, Portable Skills for Embodied Agents, and What Recursive Self-Improvement Adds
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 18, 08:46 AM

  Embodied AI lacks the data-rich pretraining that gave language models their agents and skills, so new manipulation capability costs demonstrations per task or a frontier model in every episode.

  We study a third route: a strong general model acts only as an offline author that writes, runs, diagnoses and repairs a manipulation skill in simulation until it passes an exam on layouts drawn after freezing; the frozen skill then runs as code alone or under a cheap caller that reads only its description.

Extractive summary: sentences quoted from the sources.