Skip to content

CVE-2026-33626

Server-Side Request Forgery (SSRF): Is CVE-2026-33626 real, exploitable, or a false positive? Here's the community verdict.

Community verdict: No field verdicts yet. Be the first practitioner to weigh in.

Give your verdict☆ Watch

signals

public sources

Exploited in wild
Yes
VulnCheck KEV · Apr 22, 2026
Public exploit
PoC (unverified)
Metasploit/EDB/PoC
Base severity
7.5 High
CVSS
Exploitation prob.
45%
FIRST EPSS
Weakness
CWE-918 · Server-Side Request Forgery (SSRF)
CWE

Reported exploited in the wild (VulnCheck KEV), though not in CISA KEV yet. Treat it as urgent and check your exposure.

cisa ssvc

CISA Vulnrichment · assessed

Exploitation
Public PoC
in the wild
Automatable
Yes
at scale
Technical impact
Partial
on success

CISA decision, by how critical the affected system is to you:

  • low impact → Track
  • medium impact → Track
  • high impact → Attend

Track: normal patch cycle · Track*: watch closely · Attend: patch sooner · Act: patch now. About SSVC ↗

public exploits

links to sources — we don’t host code
GitHub PoCunverified

Unverified proof-of-concept code has been published. It may or may not be functional — assess before relying on it.

affected packages

GHSA-6w67-hwm5-92mq · GitHub rates it high
  • PyPIlmdeploy
    Vulnerable
    ≤ 0.12.2
    Fixed in
    no patched version listed

Source: GitHub Advisory Database (CC-BY 4.0), via OSV.dev. Check your lockfile version against the ranges above.

baseline read

auto · not a community verdict

Low signal — verdict needed

Few public signals point to active risk. Whether a scanner hit here is a true or false positive depends on your version and config — community verdicts decide.

Based on CVSS · FIRST EPSS

Confirm or dispute →
Affected:Internlm

CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N

LMDeploy is a toolkit for compressing, deploying, and serving large language models. Versions prior to 0.12.3 have a Server-Side Request Forgery (SSRF) vulnerability in LMDeploy's vision-language module. The `load_image()` function in `lmdeploy/vl/utils.py` fetches arbitrary URLs without validating internal/private IP addresses, allowing attackers to access cloud metadata services, internal networks, and sensitive resources. Version 0.12.3 patches the issue.

References

NVD MITRE CVE.org VulnCheck KEV

Published

Embed this verdict
TruePositive verdict for CVE-2026-33626
Markdown
[![TruePositive verdict](https://www.truepositive.app/cve/CVE-2026-33626/badge.svg)](https://www.truepositive.app/cve/CVE-2026-33626)
HTML
<a href="https://www.truepositive.app/cve/CVE-2026-33626"><img src="https://www.truepositive.app/cve/CVE-2026-33626/badge.svg" alt="TruePositive verdict for CVE-2026-33626"></a>

Live badge that updates automatically as the community verdict changes.

timeline

  1. Added to VulnCheck KEV · reported exploited in the wild
  2. CISA SSVC assessment · exploitation: poc
  3. Published by NVD

Score and exploit changes tracked since .

Community ground truth

Be the first practitioner to weigh in

So far this is only TruePositive's editorial baseline from public sources. Add your real-world verdict below — it becomes the signal the next person triaging this relies on.

🥇 The first 50 practitioners to contribute earn a Founding Contributor badge.

In your experience, is this finding real and exploitable?

awaiting field verdicts
Real, but not a risk here
Not a real issue

Curated baseline: A curated baseline from public sources, shown separately from community verdicts.

No account needed. Anonymous verdicts post as an unverified signal. Log in to make yours verified and earn reputation.

Field notes & remediation

Verdicts are the quick signal. Notes are the evidence and fixes behind them.

  • 0
    Field note · TruePositive EditorialCurated

    Proof-of-concept code is on GitHub. Exploitation has been demonstrated, though it may need adapting to a real target. FIRST EPSS estimates about a 45% chance of exploitation in the next 30 days, which is high relative to most CVEs. It is reachable over the network with no authentication and no user interaction, which is the most dangerous profile.

Add a field note or remediationoptional
Note type

What are you adding?

Markdown supported · minimum 20 characters.

Related CVEs

Same weakness: CWE-918 · Server-Side Request Forgery (SSRF).