CyberRota Analysis
AI-GeneratedThe vulnerability affects the vLLM inference engine, where the Harmony tool fails to preserve the cache_salt value during "POST /v1/responses" requests, allowing an authenticated tenant to exploit cached data and potentially infer information about other tenants' interactions. This could lead to a breach of tenant isolation in environments with prefix caching enabled, which is the default setting. Organizations using vLLM versions prior to 0.30.0 should prioritize upgrading to mitigate risks associated with this issue.
Public Exploit Signal
A public exploit, PoC, GitHub repository or Metasploit reference was detected for this CVE.
Note: these links are listed for security research and verification purposes only.
Original NVD Description
vLLM is an inference and serving engine for large language models. Prior to 0.30.0, Harmony tool continuations submitted through "POST /v1/responses" requests rebuild the next-turn engine input without preserving the cache_salt value, placing the continuation prefix in the global unsalted cache namespace even when the caller enabled salting. On deployments with prefix caching enabled, which is the default, an authenticated tenant who can reconstruct a victim's low-entropy post-tool history can submit the same continuation and use the cached_tokens_per_turn count to determine whether the prefix was previously processed, defeating the intended tenant isolation of salted prefix caching. This issue is fixed in version 0.30.0.
Related CVEs
Other vulnerabilities affecting the same vendor(s)