Users report Claude Opus 5.5 reverse-engineering ability was quietly nerfed

Users say Claude Opus 5.5 now aborts reverse-engineering requests mid-sentence, citing a safety check. The reports coincide with a string of Claude product updates on October 21 that Anthropic did not document as a model change.

• Reverse-engineering prompts now stop mid-response due to a safety guardrail
• The reports follow Claude product updates on October 21 with no patch notes
• Some users say the safety classifier was always inconsistent
• The change is linked to a wave of AI game decompilations and device modding