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.
{
"cna_assigner": "GitHub_M",
"cwe_ids": [
"CWE-639"
],
"osv_generated_from": "https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/105xxx/CVE-2026-105755.json",
"unresolved_ranges": [
{
"extracted_events": [
{
"fixed": "030.0"
}
],
"source": "AFFECTED_FIELD"
}
]
}