nday
CVE-2026-100651

vllm before 0.29.0 Denial of Service via Decoder Prompt Length Bypass

vllm-project / vllm

Disclosed . nday has no fix record for it yet; that does not mean no patch exists. None of the four PoC sources lists it. Not in CISA KEV.

tags weakness typeresource exhaustion accessnetworkno user interaction productvLLM product typeAI / ML
why these tags

nday derives these tags by fixed rules from the facts below; NVD, the CNA and other sources do not assign them. Vocabulary and rules: ~/tags.

resource exhaustion
CWE-400, stated by the CNA VulnCheck and NVD, republishing [email protected]
network
AV:N in CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N (CVSS 4.0) from the CNA VulnCheck
no user interaction
UI:N in CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N (CVSS 4.0) from the CNA VulnCheck
vLLM
affected products in the CVE record: vllm-project · vllmCPE names NVD's analysts give as vulnerable: a:vllm:vllm
AI / ML
affected products in the CVE record: vllm-project · vllmCPE names NVD's analysts give as vulnerable: a:vllm:vllmthe product type of vLLM in nday's product registry
stage
DIS
disclosed
cvss
7.1 HIGH v4.0

Lifecycle

x axis UTC; dot = exact minute, wide bar = date only, hollow = unknown or inferred.
time unknown
DISdisclosed
09-26T13:23Z
PATpatch
no record yet
ANAanalysis
no record yet
POCPoC
no record yet
ITWexploited
no record yet
KEVKEV
no record yet

Summary

vLLM before 0.29.0 fails to enforce decoder prompt-length validation on the disaggregated serving endpoint /inference/v1/generate. When the request contains a 'features' (multimodal) payload, vllm/entrypoints/serve/disagg/serving.py builds a multimodal EngineInput directly from the caller-supplied token_ids, and GenerateRequest.token_ids (vllm/entrypoints/serve/disagg/protocol.py) is not checked against model_config.max_model_len. For multimodal processors that report skip_prompt_length_check=True (for example Nemotron Parse, Whisper, and FireRedLID), InputProcessor._validate_prompt_len() returns immediately for both encoder and decoder prompts, so an overlong prompt becomes an EngineCoreRequest and reaches the worker input-batch copy into a fixed max_model_len-wide NumPy row. A client able to reach the endpoint on an affected model configuration can therefore submit an overlong token_ids list to trigger a worker failure and denial of service. Fixed in 0.29.0.

Original text of the CVE record (cvelist), shown as published.

Exploit conditions

What an attack needs, read from the CVSS vector.
Attack vector AV:N
Network: the attacker only needs to reach it over the network
Attack complexity AC:L
Low: no target-specific security measures to get around
Attack requirements AT:N
None: success does not depend on how the target is deployed or running
Privileges required PR:L
Low: ordinary user privileges
User interaction UI:N
None: no action by anyone other than the attacker

CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N

CVSS 4.0 vector from the CVE record, scored by the CNA VulnCheck.

Event log

+DISdisclosed

CVE record published by VulnCheck. CVSS 7.1, high. CISA's SSVC assessment rates exploitation as poc.

src cvelist · nday , bulk import · raw
{
  "id": 458865,
  "type": "disclosed",
  "source": "cvelist",
  "url": "https://www.cve.org/CVERecord?id=CVE-2026-100651",
  "occurred_at": "2026-09-26T13:23:21.141Z",
  "observed_at": "2026-11-13T05:40:22.207467Z",
  "is_backfill": true,
  "dedupe_key": "",
  "detail": {
    "cvss": {
      "source": "cna",
      "vector": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N",
      "version": "4.0",
      "base_score": 7.1,
      "base_severity": "HIGH"
    },
    "ssvc": {
      "timestamp": "2026-09-28T18:21:26.899493Z",
      "automatable": "no",
      "exploitation": "poc",
      "technical_impact": "partial"
    },
    "state": "PUBLISHED",
    "date_published": "2026-09-26T13:23:21.141Z",
    "assigner_short_name": "VulnCheck"
  }
}

Report an error

report an error or omission

Something on this page wrong or missing? Tell the editors; if they accept it, this page's revision log says so publicly. Upstream data is not changed.

Researcher discussion

0 researchers mentioned this in the last 7 days

Posts on X by researchers on nday's list that mention this CVE, shown as posted. nday does not check them.

nday has not collected any post on X by listed researchers that mentions this CVE.

Discussion

0 comments

No comments yet.