vLLM: Out-of-vocabulary prompt token id passes input validation and triggers a CUDA device-side assert (engine crash / DoS)

For a model whose vocabulary size is an exact multiple of the embedding padding size 64 — e.g. any Llama-2-family model, vocab 32000 — a single request carrying a pre-tokenized prompt with token id equal to the vocab size 32000 passes vLLM's input-id validation but is one past the last embedding...

GHSA-Q97C-P368-75C4 vLLM: Out-of-vocabulary prompt token id passes input validation and triggers a CUDA device-side assert (engine crash / DoS)

For a model whose vocabulary size is an exact multiple of the embedding padding size 64 — e.g. any Llama-2-family model, vocab 32000 — a single request carrying a pre-tokenized prompt with token id equal to the vocab size 32000 passes vLLM's input-id validation but is one past the last embedding...

CVE-2026-116174 vLLM 0.22.0 before 0.24.0 Denial of Service via Rust gRPC Generate

vLLM from 0.22.0 before 0.24.0 contains an improper input validation vulnerability in the Rust gRPC Generate and GenerateStream service that forwards unchecked prompt token IDs to EngineCore. Attackers can submit requests with out-of-vocabulary prompt.tokenids values to trigger a fatal embedding...

Integer overflow in TFLite

Impact An attacker can craft a TFLite model that would cause an integer overflow in embedding lookup operations:cc int embeddingsize = 1; int lookupsize = 1; for int i = 0; i data.i32i; lookupsize = dim; outputshape-datak = dim; for int i = 1; i datak = dim; Both embeddingsize and lookupsize are...

PYSEC-2026-3146 Integer overflow in TFLite

Impact An attacker can craft a TFLite model that would cause an integer overflow in embedding lookup operations: cc int embeddingsize = 1; int lookupsize = 1; for int i = 0; i data.i32i; lookupsize = dim; outputshape-datak = dim; for int i = 1; i datak = dim; Both embeddingsize and lookupsize are...

PT-2026-59763

Impact An attacker can craft a TFLite model that would cause an integer overflow in embedding lookup operations: cc int embedding size = 1; int lookup size = 1; for int i = 0; i data.i32i; lookup size = dim; output shape-datak = dim; for int i = 1; i datak = dim; Both embedding size and lookup si...

CVE-2022-23559

Tensorflow is an Open Source Machine Learning Framework. An attacker can craft a TFLite model that would cause an integer overflow in embedding lookup operations. Both embeddingsize and lookupsize are products of values provided by the user. Hence, a malicious user could trigger overflows in the...

HE-LRM: Encrypted Deep Learning Recommendation Models Using Fully Homomorphic Encryption

Fully Homomorphic Encryption FHE is an encryption scheme that not only encrypts data but also allows for computations to be applied directly on the encrypted data. While computationally expensive, FHE can enable privacy-preserving neural inference in the client-server setting: a client encrypts...

SUSE CVE-2022-23559

Tensorflow is an Open Source Machine Learning Framework. An attacker can craft a TFLite model that would cause an integer overflow in embedding lookup operations. Both embeddingsize and lookupsize are products of values provided by the user. Hence, a malicious user could trigger overflows in the...

CVE-2021-33648

When performing the inference shape operation of Affine, Concat, MatMul, ArgMinMax, EmbeddingLookup, and Gather operators, if the input shape size is 0, it will access data outside of bounds of shape which allocated from heap buffers...

Huawei MindSpore Community Buffer Error Vulnerability

Huawei MindSpore Community is an open source deep learning framework from Huawei, China. Huawei MindSpore Community suffers from an information disclosure vulnerability that stems from accessing a shape allocated from the heap buffer if the input shape size is 0 when performing inferred shape...

GHSA-98P5-X8X4-C9M5 Integer overflow in TFLite

Impact An attacker can craft a TFLite model that would cause an integer overflow in embedding lookup operations: cc int embeddingsize = 1; int lookupsize = 1; for int i = 0; i data.i32i; lookupsize = dim; outputshape-datak = dim; for int i = 1; i datak = dim; Both embeddingsize and lookupsize are...

Integer overflow in TFLite

Impact An attacker can craft a TFLite model that would cause an integer overflow in embedding lookup operations: cc int embeddingsize = 1; int lookupsize = 1; for int i = 0; i data.i32i; lookupsize = dim; outputshape-datak = dim; for int i = 1; i datak = dim; Both embeddingsize and lookupsize are...

Google Tensorflow Input Validation Error Vulnerability (CNVD-2022-09880)

Google TensorFlow is an end-to-end open source platform for machine learning from Google, Inc. Google Tensorflow is vulnerable to an input validation error that could be exploited by an attacker to build a TFLite model that leads to an integer overflow in the embedding lookup operation...

Integer overflow

Tensorflow is an Open Source Machine Learning Framework. An attacker can craft a TFLite model that would cause an integer overflow in embedding lookup operations. Both embeddingsize and lookupsize are products of values provided by the user. Hence, a malicious user could trigger overflows in the...

PYSEC-2022-123

Tensorflow is an Open Source Machine Learning Framework. An attacker can craft a TFLite model that would cause an integer overflow in embedding lookup operations. Both embeddingsize and lookupsize are products of values provided by the user. Hence, a malicious user could trigger overflows in the...

PYSEC-2022-68

Tensorflow is an Open Source Machine Learning Framework. An attacker can craft a TFLite model that would cause an integer overflow in embedding lookup operations. Both embeddingsize and lookupsize are products of values provided by the user. Hence, a malicious user could trigger overflows in the...

PYSEC-2022-68

Tensorflow is an Open Source Machine Learning Framework. An attacker can craft a TFLite model that would cause an integer overflow in embedding lookup operations. Both embeddingsize and lookupsize are products of values provided by the user. Hence, a malicious user could trigger overflows in the...

PYSEC-2022-123

Tensorflow is an Open Source Machine Learning Framework. An attacker can craft a TFLite model that would cause an integer overflow in embedding lookup operations. Both embeddingsize and lookupsize are products of values provided by the user. Hence, a malicious user could trigger overflows in the...

CVE-2022-23559 Integer overflow in TFLite

Tensorflow is an Open Source Machine Learning Framework. An attacker can craft a TFLite model that would cause an integer overflow in embedding lookup operations. Both embeddingsize and lookupsize are products of values provided by the user. Hence, a malicious user could trigger overflows in the...