{"api_version":"1","generated_at":"2026-10-06T02:03:30+00:00","cve":"CVE-2026-105755","urls":{"html":"https://cve.report/CVE-2026-105755","api":"https://cve.report/api/cve/CVE-2026-105755.json","docs":"https://cve.report/api","cve_org":"https://www.cve.org/CVERecord?id=CVE-2026-105755","nvd":"https://nvd.nist.gov/vuln/detail/CVE-2026-105755"},"summary":{"title":"vLLM: Flash late-interaction scoring caches query embeddings under a caller-controlled request id — cross-request integrity break and induced errors on `/score` and `/rerank`","description":"vLLM is an inference and serving engine for large language models. Prior to 0.30.0, flash late-interaction scoring at the /score and /rerank endpoints derives each worker's query_key value from the caller-controlled X-Request-Id header. A concurrent request that reuses a victim's identifier can overwrite the cached query embedding so the victim's documents are scored against the attacker's query, and shared use counters can also cause a late-interaction cache-miss error. This issue is fixed in version 0.30.0.","state":"PUBLISHED","assigner":"GitHub_M","published_at":"2026-10-05 23:17:02","updated_at":"2026-10-05 23:17:02"},"problem_types":["CWE-639","CWE-639 CWE-639: Authorization Bypass Through User-Controlled Key"],"metrics":[{"version":"3.1","source":"security-advisories@github.com","type":"Secondary","score":"4.2","severity":"MEDIUM","vector":"CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:L","data":{"version":"3.1","vectorString":"CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:L","baseScore":4.2,"baseSeverity":"MEDIUM","attackVector":"NETWORK","attackComplexity":"HIGH","privilegesRequired":"LOW","userInteraction":"NONE","scope":"UNCHANGED","confidentialityImpact":"NONE","integrityImpact":"LOW","availabilityImpact":"LOW"}},{"version":"3.1","source":"CNA","type":"DECLARED","score":"4.2","severity":"MEDIUM","vector":"CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:L","data":{"attackComplexity":"HIGH","attackVector":"NETWORK","availabilityImpact":"LOW","baseScore":4.2,"baseSeverity":"MEDIUM","confidentialityImpact":"NONE","integrityImpact":"LOW","privilegesRequired":"LOW","scope":"UNCHANGED","userInteraction":"NONE","vectorString":"CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:L","version":"3.1"}}],"references":[{"url":"https://github.com/vllm-project/vllm/releases/tag/v0.30.0","name":"https://github.com/vllm-project/vllm/releases/tag/v0.30.0","refsource":"security-advisories@github.com","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://github.com/vllm-project/vllm/security/advisories/GHSA-2phq-3phc-84px","name":"https://github.com/vllm-project/vllm/security/advisories/GHSA-2phq-3phc-84px","refsource":"security-advisories@github.com","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://github.com/vllm-project/vllm/commit/ee17d0d869203ef9a35ad73358a4987bba14b1fc","name":"https://github.com/vllm-project/vllm/commit/ee17d0d869203ef9a35ad73358a4987bba14b1fc","refsource":"security-advisories@github.com","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://github.com/vllm-project/vllm/pull/51445","name":"https://github.com/vllm-project/vllm/pull/51445","refsource":"security-advisories@github.com","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://www.cve.org/CVERecord?id=CVE-2026-105755","name":"CVE Program record","refsource":"CVE.ORG","tags":["canonical"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-105755","name":"NVD vulnerability detail","refsource":"NVD","tags":["canonical","analysis"]}],"affected":[{"source":"CNA","vendor":"vllm-project","product":"vllm","version":"affected < 030.0","platforms":[]}],"timeline":[],"solutions":[],"workarounds":[],"exploits":[],"credits":[],"nvd_cpes":[],"vendor_comments":[],"enrichments":{"kev":null,"epss":null,"legacy_qids":[]},"source_records":{"cve_program":{"containers":{"cna":{"affected":[{"product":"vllm","vendor":"vllm-project","versions":[{"status":"affected","version":"< 030.0"}]}],"descriptions":[{"lang":"en","value":"vLLM is an inference and serving engine for large language models. Prior to 0.30.0, flash late-interaction scoring at the /score and /rerank endpoints derives each worker's query_key value from the caller-controlled X-Request-Id header. A concurrent request that reuses a victim's identifier can overwrite the cached query embedding so the victim's documents are scored against the attacker's query, and shared use counters can also cause a late-interaction cache-miss error. This issue is fixed in version 0.30.0."}],"metrics":[{"cvssV3_1":{"attackComplexity":"HIGH","attackVector":"NETWORK","availabilityImpact":"LOW","baseScore":4.2,"baseSeverity":"MEDIUM","confidentialityImpact":"NONE","integrityImpact":"LOW","privilegesRequired":"LOW","scope":"UNCHANGED","userInteraction":"NONE","vectorString":"CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:L","version":"3.1"}}],"problemTypes":[{"descriptions":[{"cweId":"CWE-639","description":"CWE-639: Authorization Bypass Through User-Controlled Key","lang":"en","type":"CWE"}]}],"providerMetadata":{"dateUpdated":"2026-10-05T22:47:54.854Z","orgId":"a0819718-46f1-4df5-94e2-005712e83aaa","shortName":"GitHub_M"},"references":[{"name":"https://github.com/vllm-project/vllm/security/advisories/GHSA-2phq-3phc-84px","tags":["x_refsource_CONFIRM"],"url":"https://github.com/vllm-project/vllm/security/advisories/GHSA-2phq-3phc-84px"},{"name":"https://github.com/vllm-project/vllm/pull/51445","tags":["x_refsource_MISC"],"url":"https://github.com/vllm-project/vllm/pull/51445"},{"name":"https://github.com/vllm-project/vllm/commit/ee17d0d869203ef9a35ad73358a4987bba14b1fc","tags":["x_refsource_MISC"],"url":"https://github.com/vllm-project/vllm/commit/ee17d0d869203ef9a35ad73358a4987bba14b1fc"},{"name":"https://github.com/vllm-project/vllm/releases/tag/v0.30.0","tags":["x_refsource_MISC"],"url":"https://github.com/vllm-project/vllm/releases/tag/v0.30.0"}],"source":{"advisory":"GHSA-2phq-3phc-84px","discovery":"UNKNOWN"},"title":"vLLM: Flash late-interaction scoring caches query embeddings under a caller-controlled request id — cross-request integrity break and induced errors on `/score` and `/rerank`"}},"cveMetadata":{"assignerOrgId":"a0819718-46f1-4df5-94e2-005712e83aaa","assignerShortName":"GitHub_M","cveId":"CVE-2026-105755","datePublished":"2026-10-05T22:47:54.854Z","dateReserved":"2026-10-05T19:11:07.947Z","dateUpdated":"2026-10-05T22:47:54.854Z","state":"PUBLISHED"},"dataType":"CVE_RECORD","dataVersion":"5.2"},"nvd":{"publishedDate":"2026-10-05 23:17:02","lastModifiedDate":"2026-10-05 23:17:02","problem_types":["CWE-639","CWE-639 CWE-639: Authorization Bypass Through User-Controlled Key"],"metrics":{"cvssMetricV31":[{"source":"security-advisories@github.com","type":"Secondary","cvssData":{"version":"3.1","vectorString":"CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:L","baseScore":4.2,"baseSeverity":"MEDIUM","attackVector":"NETWORK","attackComplexity":"HIGH","privilegesRequired":"LOW","userInteraction":"NONE","scope":"UNCHANGED","confidentialityImpact":"NONE","integrityImpact":"LOW","availabilityImpact":"LOW"},"exploitabilityScore":1.6,"impactScore":2.5}]},"configurations":[]},"legacy_mitre":{"record":{"CveYear":"2026","CveId":"105755","Ordinal":"1","Title":"vLLM: Flash late-interaction scoring caches query embeddings und","CVE":"CVE-2026-105755","Year":"2026"},"notes":[{"CveYear":"2026","CveId":"105755","Ordinal":"1","NoteData":"vLLM is an inference and serving engine for large language models. Prior to 0.30.0, flash late-interaction scoring at the /score and /rerank endpoints derives each worker's query_key value from the caller-controlled X-Request-Id header. A concurrent request that reuses a victim's identifier can overwrite the cached query embedding so the victim's documents are scored against the attacker's query, and shared use counters can also cause a late-interaction cache-miss error. This issue is fixed in version 0.30.0.","Type":"Description","Title":"vLLM: Flash late-interaction scoring caches query embeddings und"}]}}}