ACE-CaP-0 researchers reported that robot programs developed by Anthropic's Claude Opus 5.5 coding agent succeeded in 221 of 240 physical trials, or 92.1%. The evaluation covered eight tasks, with each program fixed before 30 trials and no human or coding agent directing execution. Human advice and physical demonstrations were available while the agent developed the programs. Consistency across independent development runs remains unestablished, and testing under new conditions is needed to determine broader reuse.