{"dataType": "CVE_RECORD", "dataVersion": "5.2", "cveMetadata": {"cveId": "CVE-2026-54234", "assignerOrgId": "a0819718-46f1-4df5-94e2-005712e83aaa", "state": "PUBLISHED", "assignerShortName": "GitHub_M", "dateReserved": "2026-06-12T16:25:43.084Z", "datePublished": "2026-07-06T19:49:20.481Z", "dateUpdated": "2026-07-06T19:49:20.481Z"}, "containers": {"cna": {"title": "vLLM: Remote DoS in vLLM via Invalid Recovered Token Reinjection", "problemTypes": [{"descriptions": [{"cweId": "CWE-20", "lang": "en", "description": "CWE-20: Improper Input Validation", "type": "CWE"}]}, {"descriptions": [{"cweId": "CWE-1284", "lang": "en", "description": "CWE-1284: Improper Validation of Specified Quantity in Input", "type": "CWE"}]}], "metrics": [{"cvssV3_1": {"attackComplexity": "LOW", "attackVector": "NETWORK", "availabilityImpact": "HIGH", "baseScore": 7.5, "baseSeverity": "HIGH", "confidentialityImpact": "NONE", "integrityImpact": "NONE", "privilegesRequired": "NONE", "scope": "UNCHANGED", "userInteraction": "NONE", "vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H", "version": "3.1"}}], "references": [{"name": "https://github.com/vllm-project/vllm/security/advisories/GHSA-8wr5-jm2h-8r4f", "tags": ["x_refsource_CONFIRM"], "url": "https://github.com/vllm-project/vllm/security/advisories/GHSA-8wr5-jm2h-8r4f"}, {"name": "https://github.com/vllm-project/vllm/pull/44744", "tags": ["x_refsource_MISC"], "url": "https://github.com/vllm-project/vllm/pull/44744"}, {"name": "https://github.com/vllm-project/vllm/commit/8a5cf1ccd65e8ac7635c402c1ec0b08988bc26ca", "tags": ["x_refsource_MISC"], "url": "https://github.com/vllm-project/vllm/commit/8a5cf1ccd65e8ac7635c402c1ec0b08988bc26ca"}], "affected": [{"vendor": "vllm-project", "product": "vllm", "versions": [{"version": "< 0.24.0", "status": "affected"}]}], "providerMetadata": {"orgId": "a0819718-46f1-4df5-94e2-005712e83aaa", "shortName": "GitHub_M", "dateUpdated": "2026-07-06T19:49:20.481Z"}, "descriptions": [{"lang": "en", "value": "vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Prior to 0.24.0, a frontend-legal multi-request speculative decoding workload can cause the rejection sampler to produce a recovered token equal to the model vocabulary size boundary value, which is then converted to negative one when the engine selects the next live token for a request and is written back into the drafter's input ids; that out-of-vocabulary value is later consumed by the model's embedding and attention path and crashes the engine worker with a GPU device-side assertion. The same triggering request sequence is reachable through the public gRPC Generate and Abort endpoints, so a remote client that can send generation requests can crash the shared engine worker, aborting concurrent requests and causing a service-wide denial of service for other clients of the deployment until the worker is restarted. This issue is fixed in version 0.24.0."}], "source": {"advisory": "GHSA-8wr5-jm2h-8r4f", "discovery": "UNKNOWN"}}}}

{}

{}

{}

{}