CVE-2026-100652

vLLM versions 0.22.0 through 0.23.0 fail to validate stoptokenids against vocabulary bounds in Rust HTTP and gRPC frontends, allowing out-of-vocabulary token IDs to reach MinTokensLogitsProcessor. Attackers can submit requests with mintokens greater than zero and out-of-vocabulary stoptokenids to...

PT-2026-99325

Name of the Vulnerable Software and Affected Versions vLLM versions prior to 0.29.0 Description The software fails to validate if the values provided in stop token ids are within the model vocabulary or logits range, checking only that they are integers. When min tokens is greater than 0, these I...

CVE-2026-93592

A flaw was found in vLLM. An unauthenticated attacker can exploit an input validation vulnerability by submitting negative token IDs to the /v1/embeddings and /pooling endpoints. This action triggers a CUDA device-side assertion, which poisons the GPU context and causes the engine to crash...

CVE-2026-93841 vLLM through 0.29.0 Adjacent-Request Sampler State Corruption via Unvalidated Prompt Token IDs

vLLM through 0.29.0 contains a memory corruption vulnerability in the Triton bincountkernel where prompt token IDs index the penalty prompt-presence bitset without bounds checking against vocabulary size. Attackers can submit multimodal audio requests with tokens equal to vocabulary size, causing...

CVE-2026-93592

vLLM versions before 0.28.0 fail to validate the lower bound of token IDs in the /v1/embeddings and /pooling endpoints, allowing unauthenticated attackers to crash the engine by submitting negative token IDs. A single request with a negative token ID triggers a CUDA device-side assertion that...

EUVD-2026-82911

vLLM versions before 0.28.0 fail to validate the lower bound of token IDs in the /v1/embeddings and /pooling endpoints, allowing unauthenticated attackers to crash the engine by submitting negative token IDs. A single request with a negative token ID triggers a CUDA device-side assertion that...

CVE-2026-93592: Improper Validation of Array Index

vLLM versions before 0.28.0 fail to validate the lower bound of token IDs in the /v1/embeddings and /pooling endpoints, allowing unauthenticated attackers to crash the engine by submitting negative token IDs. A single request with a negative token ID triggers a CUDA device-side assertion that...

vLLM: Derender endpoints decode caller-supplied GenerateResponse token IDs without output bounds

SummaryThe /v1/completions/derender and /v1/chat/completions/derender endpoints accept caller-supplied GenerateResponse objects and postprocess every nested choices.tokenids list directly. Unlike the normal render/generate path, derender does not enforce model context length, resolved maxtokens,...

PYSEC-2026-3936 vLLM: Derender endpoints decode caller-supplied GenerateResponse token IDs without output bounds

Summary The /v1/completions/derender and /v1/chat/completions/derender endpoints accept caller-supplied GenerateResponse objects and postprocess every nested choices.tokenids list directly. Unlike the normal render/generate path, derender does not enforce model context length, resolved maxtokens,...

vLLM denial of service via prompt embeds on M-RoPE models

Summary Short summary of the problem. Make the impact and severity as clear as possible. For example: An unsafe deserialization vulnerability allows any unauthenticated user to execute arbitrary code on the server. Sending a pure prompt embeds payload in a /v1/completions request with a model usi...

EUVD-2019-2206

Malware in sbrugna...

Race condition vulnerability in positionManager minting function

Lines of code Vulnerability details Impact race condition vulnerability can result in the issuance of duplicate token IDs. When multiple transactions are executed in quick succession attempting to mint tokens, they may end up being assigned the same ID due to a shared counter or variable used to...

Upgraded Q -> M from #445 [1674423223201]

Judge has assessed an item in Issue 445 as M risk. The relevant finding follows: 5 Function crossChain in GovNFT should have limit for maximum tokens allowed to be transferred, because of gas limit in the dest chain. if a user transferred a lot of tokens because there was two loop inside each oth...

ownerMintUsingTokenId can brick the whole contract

Lines of code Vulnerability details Impact With the function ownerMintUsingTokenId, it is possible for the owner to mint a token with an arbitrary token ID. However, this can brick the whole contract and cause a situation where no more mints / buys are possible. This happens when a token ID is...

ETH can get stuck (and possibly be stolen as well).

Lines of code Vulnerability details Impact The Turnstile contract contains a distributeFees function which can only be called by owner to assign and distribute fee for a tokenId. However the function does not validates the tokenId input value. So it is possible for the owner to provide any uint25...

Non unique token ID might lead to collusion when merging

Handle kenzo Vulnerability details Token IDs are defined as concatenation of points, total supply + 1. The total supply can decrease when merging. This means that the contract might try to mint a token with an ID which already exists. Impact Under specific circumstances, users won't be able to lo...

Cross-site Scripting (XSS)

pki-core is vulnerable to cross-site scripting XSS. A remote attacker is able to inject and execute arbitrary Javascript code in a user's browser via the Token IDs from the Activity page...

CVE-2019-10178: Improper Neutralization of Input During Web Page Generation

It was found that the Token Processing Service TPS did not properly sanitize the Token IDs from the "Activity" page, enabling a Stored Cross Site Scripting XSS vulnerability. An unauthenticated attacker could trick an authenticated victim into creating a specially crafted activity, which would...