ai_security▌

← feed

NOTABLE

vLLM advisory: thinking-budget state aliases onto a co-batched request, forcing reasoning-end tokens cross-session

On the legacy V1 model runner, a request using the public thinking_token_budget parameter has its budget-reached sparse state aliased onto a co-batched request during a mixed SWAP, forcing the configured reasoning-end token sequence into the other request's output (CWE-488, exposure of a data element to the wrong session). The repo security advisory (medium, published Nov 9 17:30 UTC) covers v0.21.0 through v0.27.1, fixed in v0.28.0. First ingestion.

released 2026-11-09 · ingested 2026-11-11T02:40Z

Sources

Primary source

Cite this

APA

AI Security Index. (2026). vLLM advisory: thinking-budget state aliases onto a co-batched request, forcing reasoning-end tokens cross-session. AI Security Index. https://github.com/vllm-project/vllm/security/advisories/GHSA-fpch-3xrf-7q96

BibTeX

@misc{aisec-2026-11-09-vllm-thinking-budget-state-aliasing,
  title  = {vLLM advisory: thinking-budget state aliases onto a co-batched request, forcing reasoning-end tokens cross-session},
  year   = {2026},
  howpublished = {\url{https://github.com/vllm-project/vllm/security/advisories/GHSA-fpch-3xrf-7q96}},
  note   = {AI Security Index item 2026-11-09-vllm-thinking-budget-state-aliasing}
}