# CVE-2026-105754: vLLM inference engine input validation bypass in scale-out

Source: https://junglewise.ai/threats/cve-2026-105754-vllm-is-an-inference-and-serving-engine-for-large-language  
Publisher: Junglewise Threat Intelligence, https://junglewise.ai/threats  
Published: 2026-10-05. Updated: 2026-10-05  
License: CC BY 4.0 with a link back

CVE-2026-105754 · Severity: medium · CVSS 6.5 · Published 2026-10-05

vLLM is an inference engine for large language models that can be deployed in a distributed scale-out configuration. An attacker can send malicious requests to the /inference/v1/generate endpoint with forged tensor data, cache identifiers, or placeholder ranges to crash the shared engine, poison cached encoder state, or alter request processing semantics. This could lead to denial of service or unauthorized access to cached model processing state across requests.

Technologies: [vllm (PyPI)](https://junglewise.ai/threats/technologies/pypi-vllm), [Vllm](https://junglewise.ai/threats/technologies/vllm). Vendors: [PyPI](https://junglewise.ai/threats/vendors/pypi), [Vllm](https://junglewise.ai/threats/vendors/vllm).

## Quick answers

**Is CVE-2026-105754 exploited in the wild?** No. Junglewise Threat Intelligence has no report of CVE-2026-105754 being exploited in the wild.

**How severe is CVE-2026-105754?** Medium severity, with a CVSS score of 6.5, published 5 October 2026.

**Which products and versions are affected?** vllm (PyPI) < 0.30.0, ECOSYSTEM:introduced=0,fixed=0.30.0; Vllm before 0.30.0.

**Is there a fix for CVE-2026-105754?** Yes. vllm (PyPI): upgrade to 0.30.0 or later. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.

**How is CVE-2026-105754 attacked?** It is attacked over the network, in a software library.

**Am I affected by CVE-2026-105754?** Check whether you run vllm (PyPI) or Vllm at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

## How to fix

- vllm (PyPI): upgrade to 0.30.0 or later

  ```sh
  pip install "vllm>=0.30.0"
  ```


## Technical details

The /inference/v1/generate endpoint in vLLM's disaggregated scale-out path accepts caller-supplied multimodal data in features.kwargs_data, features.mm_hashes, and features.mm_placeholders fields without proper validation or rebinding to the active model renderer contract. An unauthenticated network attacker can supply forged grid geometry, field types, or non-positive placeholder lengths to crash the EngineCore; known content hashes can be used to poison or retrieve cross-request encoder-cache state; and missing sparse placeholder masks can alter transport semantics. The vulnerability was patched by adding input validation before engine handoff.

## Affected products

- vLLM vLLM before 0.30.0

## Identifiers

- CVE-2026-105754
- GHSA-ph72-cqr5-qpp7
- PYSEC-2026-4309

## Timeline

- 2026-10-05: disclosed

## References

- [https://github.com/vllm-project/vllm/commit/1970f3ed4be7fa8620e4ddc4a12c36a8384cfc27](https://github.com/vllm-project/vllm/commit/1970f3ed4be7fa8620e4ddc4a12c36a8384cfc27)
- [https://github.com/vllm-project/vllm/pull/51898](https://github.com/vllm-project/vllm/pull/51898)

## Related threats

- [vLLM chat template resource exhaustion DoS](https://junglewise.ai/threats/cve-2026-115918-vllm-is-an-inference-and-serving-engine-for-large-language)
- [vLLM: Permissive default CORS origins allow cross-origin requests to the vLLM HTTP server](https://junglewise.ai/threats/cve-2026-115917-vllm-is-an-inference-and-serving-engine-for-large-language)
- [vLLM off-by-one error in tokenizer max_token_id validation](https://junglewise.ai/threats/cve-2026-115916-vllm-is-an-inference-and-serving-engine-for-large-language)
- [vLLM image format validation bypass in media processing](https://junglewise.ai/threats/cve-2026-115915-vllm-is-an-inference-and-serving-engine-for-large-language)
- [vLLM denial of service in DeepSeek-OCR image processing](https://junglewise.ai/threats/cve-2026-115914-vllm-is-an-inference-and-serving-engine-for-large-language)