CVE-2026-105755: vLLM: Flash late-interaction scoring caches query embeddings under a caller-controlled request id — cross-request integrity break and induced errors on…
Noroxi assessment
Priority: Monitor (action score 16/100)
- The sources we track (CISA KEV, Exploit-DB, Metasploit, Nuclei, GitHub) show no public exploit code or exploitation in the wild.
- Probability of exploitation in the next 30 days: 0.20% (FIRST EPSS).
- Anyone who can reach it over the internet can trigger it. A low-privileged account is enough. No user action is required. Special conditions such as timing or configuration are required.
- Fixed version: vllm-project vllm 0.30.0.
Low priority for now. Upgrade to vllm-project vllm 0.30.0 or later.
Generated automatically from the record's data: CVSS (NVD and CNA), CISA KEV and SSVC, FIRST EPSS, exploit sources, fixed versions. How we calculate it
Primary sources: NVD · CVE.org · FIRST EPSS
vLLM is an inference and serving engine for large language models. Prior to 0.30.0, flash late-interaction scoring at the /score and /rerank endpoints derives each worker's query_key value from the caller-controlled X-Request-Id header. A concurrent request that reuses a victim's identifier can overwrite the cached query embedding so the victim's documents are scored against the attacker's query, and shared use counters can also cause a late-interaction cache-miss error. This issue is fixed in version 0.30.0.
- Published
- Oct 5, 2026
- Updated
- Oct 8, 2026
- EPSS
- 0.2% · 10th percentile
- CWE
- CWE-639
Report tools
Action score
16
Low priority for now.
- CVSS
- 16 / 40 · 4.2 / 10
- CISA KEV
- 0 / 30 · Not listed
- EPSS
- 0 / 30 · 0.2%
CISA SSVC decision
- Exploitation
- none
- Automatable
- no
- Technical impact
- partial
Vulnrichment: CISA's decision-tree inputs.
CNA vs NVD score
- NVD
- —
- CNA · GitHub_M
- 4.2
- NVD has not scored this yet; the score shown is the CNA’s.
The score given by the assigning authority versus NVD’s independent score. A gap means the severity is contested.
Get told when this record changes
- Not in KEV: no record of exploitation in the wild
- No public exploit code on record
- A fix has been released
If any of these changes, we'll tell you in the daily email digest.
Is your version affected? Check itNoroxi analysis
What does it mean in your environment?
The assessment above is generated automatically from the record's data; it doesn't know your environment. Our team verifies on your systems whether this vulnerability is actually exploitable and which mitigation is enough.
We use this product, ask for helpAffected systems
Is your version affected?
Pick the product and type the version you run. It's compared with this record's version ranges in your browser; the version isn't sent anywhere.
| Vendor | Product | CPE |
|---|---|---|
| vllm | vllm | cpe:2.3:a:vllm:vllm |
Versions reported by the vendor
Affected version ranges reported by the assigning authority (GitHub_M). Independent of NVD's CPE analysis and usually ahead of it.
vllm-project vllm
- < 030.0affected
Affected versions
NVD version ranges (for catalog products). Add the product to your stack with a version and matching uses these.
- vllm vllmbefore 0.30.0
Package-level exposure
OSV and GitHub Advisory data: ecosystem, package and range. SBOM matching uses this table.
| Ecosystem | Package | Affected range | Fix |
|---|---|---|---|
| PyPI | vllm | before 0.30.0 | 0.30.0 |
Same product
vllm: all recordsOther highest-scoring records for the same primary product.
- CVE-2026-22778vLLM leaks a heap address when PIL throws an error60This week
- CVE-2025-32444vLLM Vulnerable to Remote Code Execution via Mooncake Integration40Plan
- CVE-2024-11041Remote Code Execution in vllm-project/vllm40Plan
Published before and after this one in the same product
- ← EarlierCVE-2026-105753Oct 5, 2026 · vLLM: Mirrored multimodal IPC caches desync after a rejected request — a later request reusing the same media hash trips a receiver assertion in the engine…26Monitor
- ← EarlierCVE-2026-105754Oct 5, 2026 · vLLM: Scale-out disaggregated multimodal transport trusts caller-supplied features26Monitor
- Later →CVE-2026-105756Oct 5, 2026 · vLLM: Loose `cache_salt` validation lets a single request kill EngineCore on LMCache-MP deployments — uncaught downstream `ValueError` denial of service26Monitor
Remediation
Which version to upgrade to
Fix versions compiled from the vendor, package registries and Microsoft. Verify the vendor's note before upgrading.
| Product / package | Fixed version | Source |
|---|---|---|
| vllm-project vllm | 0.30.0 · description | Vendor (CNA) |
| PyPI:vllm | 0.30.0 | Package registry (OSV) |
Exploit status
No known public exploit
No public exploit has been observed yet. That doesn't mean you're safe, only that the bar is a little higher.
Research context
For pentesters and researchers: attack profile, score disagreement, timeline, patch commits, credits, variant and chain candidates, bug bounty scope. All derived from existing data; no exploit code.
Timeline
From publication to today: proof of concept, Metasploit module, CISA KEV and fix record. Dates are as reported by the sources.
- Published
- Fix recordedsame day
- TodayNov 13, 2026+39 d
FIRST EPSS daily score; only changes of 0.01 or more are recorded (step chart).
Patch and commit links
Commit, PR and diff links among the references. A starting point for patch-diffing and variant hunting; fixes, not exploits.
Not on record for this CVE: researcher credit · variant or chain candidate · public bounty program.
Technical details
Attack conditions
- Anyone who can reach it over the internet can trigger it.
- A low-privileged account is enough.
- No user action is required.
- Special conditions such as timing or configuration are required.
If successful
- Confidentiality
- none
- Integrity
- low · data or configuration can be modified
- Availability
- low · the service can be disrupted
- Attack vector
- Network
- Attack complexity
- High
- Privileges required
- Low
- User interaction
- None
- Scope
- Unchanged
- Confidentiality impact
- None
- Integrity impact
- Low
- Availability impact
- Low
Weakness class (CWE)
CWE-639 · Authorization Bypass Through User-Controlled KeyCVSS vector
CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:L
nvd-secondary
Root cause
The endpoint fetches the record by the ID sent in the request, but never checks whether the record belongs to the requesting user.
The bug in code
The whole CWE-639 classA representative example of this vulnerability class. Not the vendor's source code; it shows the faulty pattern and its fix.
Vulnerable
const invoice = await Invoice.findById(req.params.id);
res.json(invoice);Fixed
const invoice = await Invoice.findOne({
id: req.params.id,
ownerId: req.user.id,
});
if (!invoice) return res.sendStatus(404);
res.json(invoice);Attack context
MITRE CAPEC attack patterns and ATT&CK techniques for this weakness class (CWE). A starting point for detection rules and threat hunting.
MITRE has no CAPEC/ATT&CK mapping for this CWE.
Change log
- SSVC exploitation— → none
- Exploit maturity— → none
- Fix✗ → ✓
For records you follow, these changes also arrive as notifications. →
References
Show all references (3 more)Show less
Vendor advisories and official records. Exploit/PoC links are deliberately left out.