主流大语言模型提供商如今都会隐藏模型逐步推理的过程,即思维链,以保护知识产权并限制信息泄露。提供商并非将这些推理轨迹存储在服务端,而是将其以加密文本块的形式返回给客户端,客户端在后续每次请求时再将这些文本块传回。在既有研究的基础上,我们发现了一个架构层面的漏洞:这些加密文本块在同一提供商生态内的不同会话、不同用户乃至不同模型之间完全兼容且可互换。
我们利用这一兼容性开发出了一种可扩展的解密越狱攻击。通过将某个模型生成的加密推理轨迹注入同一提供商旗下防护较弱的模型,我们迫使该模型将其逐字解码并以明文形式输出,而无需直接越狱能力更强的目标模型。这一漏洞衍生出四种不同的攻击向量。
其一,它绕过了反蒸馏机制,使攻击者能够提取专有模型的推理过程,我们已在 Anthropic、OpenAI 和 Google 上完成了验证。其二,它可实现大规模私有数据窃取。开发者经常公开分享会话日志,却并不清楚其中加密文本块的内容。
通过解码从公开代码仓库中抓取的 315,320 个推理文本块,我们恢复了 367 条个人身份信息(PII)和 182 组凭据。其三,它会在不经意间泄露推理过程中隐藏的危险信息,即便模型最终可见的输出已安全地拒绝了恶意请求。
其四,攻击者可利用这一缺陷实施隐形提示词注入,将恶意载荷完全嵌入加密文本块中,从而污染公开的智能体部署。在遵循负责任披露流程之后,我们提出了具体的密码学与系统级缓解措施,以保障客户端侧推理的安全性。
Leading large language model providers now conceal their models' step-by-step reasoning, or chain-of-thought, to protect intellectual property and limit information leakage. Rather than storing these traces server-side, providers return them to the client as blocks of encrypted text, which the client passes back with each subsequent request. Building on prior research, we identify an architectural vulnerability: these encrypted blocks are fully compatible and interchangeable across different sessions, users, and models within a provider's ecosystem.
We exploit this compatibility to develop a scalable decryption jailbreak. By injecting an encrypted reasoning trace from a given model into a weaker, and less safeguarded model from the same provider, we force it to decode and output the trace verbatim in plaintext, without ever jailbreaking the more capable model directly. This vulnerability enables four distinct attack vectors. First, it circumvents anti-distillation mechanisms, allowing adversaries to extract a proprietary model's reasoning, as we demonstrate across Anthropic, OpenAI, and Google.
Second, it allows for large-scale private data extraction. Developers frequently share session logs publicly, unaware of contents of the encrypted blocks. By decoding 315,320 reasoning blocks scraped from public repositories, we recovered 367 Personally Identifiable Information (PII) artifacts and 182 credentials. Third, it inadvertently reveals hazardous information hidden within the reasoning process, even in cases where the model's final, visible output safely rejects a malicious request. Fourth, attackers can leverage this flaw to execute invisible prompt injections, embedding malicious payloads entirely within encrypted blocks to poison public agentic rollouts.
Following responsible disclosure, we propose concrete cryptographic and system-level mitigations to secure client-side reasoning.