vLLM: Harmony tool continuations drop `cache_salt` — restoring a cross-tenant prefix-cache membership oracle (vllm-project vllm)

CVE-2026-105752
LOW 3.1

En bref

Source
NVD
EPSS
EPSS 0.22 % · centile 12 ()
État
PUBLISHED
Publiée
Mise à jour
Attribuée par
GitHub_M
Statut NVD
Analyzed

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.

Faiblesses (CWE)

  • CWE-200 — Exposition d'informations sensibles à un acteur non autorisé
  • CWE-524 — Use of Cache Containing Sensitive Information

Références

Scores

VersionScoreVecteurÉvalué par
CVSS 3.1LOW 3.1CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:NGitHub_M
CVSS 3.1LOW 3.1CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:NNVD · security-advisories@github.com (Secondary)

Avis

Références européennes

Configurations affectées (CPE NVD)

Produits affectés

ProduitVersions
vllm-project · vllmaffected < 0.30.0

Sources

SourceLicence
CVE List (CVEProject cvelistV5)CVE Program Terms of Use
EPSS (FIRST.org)FIRST — attribution requested
ENISA EUVDENISA — reproduction authorised, source acknowledged
NVD (NIST) via fkie-cad/nvd-json-data-feedsPublic domain — this product uses data from the NVD API but is not endorsed or certified by the NVD