vLLM 0.8.3 - 0.14.0 - Information Disclosure

vLLM 0.8.3 to - 0.14.1 contains an information disclosure caused by leaking a heap address in error messages from the multimodal endpoint when processing invalid images, letting remote attackers reduce ASLR entropy, exploit requires sending invalid images. id: CVE-2026-22778 info: name: vLLM 0.8....

CVE-2026-15974

SGLang contains an SSRF and local file read in the multimodal generation endpoint /v1/chat/completions due to unsanitized imageurl, allowing access to internal metadata, secrets, and services...

EUVD-2026-51263

SGLang contains an SSRF and local file read in the multimodal generation endpoint /v1/chat/completions due to unsanitized imageurl, allowing access to internal metadata, secrets, and services...

CVE-2026-15974 CVE-2026-15974

SGLang contains an SSRF and local file read in the multimodal generation endpoint /v1/chat/completions due to unsanitized imageurl, allowing access to internal metadata, secrets, and services...

CVE-2026-15974 CVE-2026-15974

SGLang contains an SSRF and local file read in the multimodal generation endpoint /v1/chat/completions due to unsanitized imageurl, allowing access to internal metadata, secrets, and services...

CVE-2026-15974

SGLang (CVE-2026-15974) has a documented SSRF and local file-read flaw in the multimodal generation endpoint /v1/chat/completions caused by unsanitized image_url, enabling access to internal metadata, secrets, and services. The connected sources confirm the vulnerability details but do not provid...

Piggybacking on Perception: Stealthy Concurrent Audio Prompt Injections against Multimodal LLM Agents

Large Language Model LLM-driven multimodal agents are increasingly deployed to execute autonomous tasks via continuous audio interaction. While this paradigm enhances interaction naturalness, it introduces a critical yet under-explored attack surface, as audio inputs inevitably contain...

PYSEC-2026-3480 LMCache: 16-bit multimodal hash collision can poison KV cache entries

A flaw has been found in LMCache up to 0.4.6. This affects the function hexhashtoint16 of the file lmcache/integration/vllm/utils.py of the component KV Cache Handler. Executing a manipulation can lead to use of weak hash. The attack needs to be launched locally. The attack requires a high level ...

EUVD-2026-44953

text-generation-inference through 3.3.7 contains a server-side request forgery SSRF vulnerability in the OpenAI-compatible multimodal chat completions endpoint that allows unauthenticated network attackers to coerce the server into issuing arbitrary HTTP GET requests by supplying a crafted imageu...

CVE-2026-63086 text-generation-inference 3.3.7 SSRF via fetch_image in multimodal chat completions

text-generation-inference through 3.3.7 contains a server-side request forgery SSRF vulnerability in the OpenAI-compatible multimodal chat completions endpoint that allows unauthenticated network attackers to coerce the server into issuing arbitrary HTTP GET requests by supplying a crafted imageu...

CVE-2026-63086

The CVE-2026-63086 entry concerns text-generation-inference 3.3.7, with an SSRF flaw in the OpenAI-compatible multimodal chat completions endpoint. The fetch_image function in router/src/validation.rs does not validate addresses (private, loopback, link-local, or cloud metadata). The HTTP client ...

CVE-2026-24234

NVIDIA TensorRT-LLM for Linux contains a vulnerability in the multimodal media fetching functions, where a network-accessible attacker could cause server-side request forgery. A successful exploit of this vulnerability might lead to denial of service and information disclosure...

CVE-2026-24234

NVIDIA TensorRT-LLM for Linux is affected by CVE-2026-24234, a vulnerability in the multimodal media fetching functions that could allow a network-accessible attacker to trigger server-side request forgery, potentially leading to denial of service and information disclosure. The NVIDIA security b...

CVE-2026-24234

NVIDIA TensorRT-LLM for Linux contains a vulnerability in the multimodal media fetching functions, where a network-accessible attacker could cause server-side request forgery. A successful exploit of this vulnerability might lead to denial of service and information disclosure...

PYSEC-2026-3409 vLLM Vulnerable to Remote DoS via Special-Token Placeholders

Summary This report explains a Token Injection vulnerability in vLLM’s multimodal processing. Unauthenticated, text-only prompts that spell special tokens are interpreted as control. Image and video placeholder sequences supplied without matching data cause vLLM to index into empty grids during...

PYSEC-2026-3047 RAGAS has an Arbitrary File Read vulnerability

An Arbitrary File Read vulnerability exists in the ImageTextPromptValue class in Exploding Gradients RAGAS v0.2.3 to v0.2.14. The vulnerability stems from improper validation and sanitization of URLs supplied in the retrievedcontexts parameter when handling multimodal inputs...

PYSEC-2026-2019 vLLM vulnerable to DoS with incorrect shape of multimodal embedding inputs

Summary Users can crash the vLLM engine serving multimodal models by passing multimodal embedding inputs with correct ndim but incorrect shape e.g. hidden dimension is wrong, regardless of whether the model is intended to support such inputs as defined in the Supported Models page. The issue has...

PYSEC-2026-2022 phi4mm: Quadratic Time Complexity in Input Token Processing​ leads to denial of service

Summary A critical performance vulnerability has been identified in the input preprocessing logic of the multimodal tokenizer. The code dynamically replaces placeholder tokens e.g., , with repeated tokens based on precomputed lengths. Due to ​​inefficient list concatenation operations​​, the...

PYSEC-2026-539 SGLang's multimodal generation module is vulnerable to unauthenticated remote code execution through the ZMQ broker

SGLang's multimodal generation module is vulnerable to unauthenticated remote code execution through the ZMQ broker, which deserializes untrusted data using pickle.loads without authentication...

CVE-2026-56340

vLLM versions = 0.10.2 and 0.13.0 are missing sparse tensor validation in multimodal embeddings processing. Because PyTorch disables sparse tensor invariant checks by default, an attacker can submit crafted embedding requests with malformed negative or out-of-bounds tensor indices, when the...