Closed-source reasoning chains extracted at scale; Coze CLI hijacks AI tools

2026-08-11 群聊日报

The big one today: researchers extracted hidden reasoning chains from Anthropic, OpenAI, and Google models at scale. The trick is absurdly simple—take Opus 4.8's encrypted CoT and feed it to Haiku 4.5, which decodes it verbatim. All three API families were broken, and decoding 10K trajectories costs about $720. A separate paper shows you can even reverse-engineer reasoning from public outputs alone using a 1.5B-param model. Separately, Coze CLI was caught silently scanning local Codex and Claude Code directories and injecting its own skills into workflows. On the engineering side, the group discussed how prompt debt now rivals traditional code debt—old rules pile up, evals lag behind model iterations, and nobody dares delete anything.

Why it matters: Strong cross-source cluster signal (chat digest + original paper + study notes). First systematic validation that encrypted CoT from three major vendors is cross-model decodable, with concrete $720/10k cost. All three HKR axes hit, but the source is a secondary digest rather t...