MediumVulnerability
GHSA-v359-jj2v-j536: vLLM has SSRF Protection Bypass
- Identifiers
- CVE-2026-25960GHSA-v359-jj2v-j536
- Published
- Record updated
- Affected
- vllm >= 0.15.1, < 0.17.0
- Fixed in
- 0.17.0
- released March 6, 2026, 2 days before disclosure
- Known exploitation
- Not listed in the CISA Known Exploited Vulnerabilities catalog at the last check.
- EPSS
- 0.6%
Summary
The SSRF protection fix in vLLM's `load_from_url_async` method, in `vllm/connections.py`, can be bypassed. The fix validates URLs with `urllib3.util.parse_url()`, but the actual requests use `aiohttp`, which relies on the `yarl` library. The two parsers handle backslash characters differently. A URL such as `https://httpbin.org\@evil.com/` passes validation with host `httpbin.org` but is requested from `evil.com`, which lets an attacker bypass the hostname allowlist and reach arbitrary internal or external services.
Mitigation
Fixed by https://github.com/vllm-project/vllm/pull/34743
Affected packages in the Exposure Registry
Matched by package name and ecosystem. Each entry shows whether the package delegates to a language model and how many tracked packages depend on it.
- vllmPyPILLM dependency since 2023-08-25 · 15 tracked dependents
Topics
Related items
- MediumCVE-2026-121936: easyVoice server-side request forgery through openaiBaseUrl in /api/v1/ttsSimilar attack · NVD/CVE Database
- MediumCVE-2026-118628: MLflow get-artifacts server-side request forgery information disclosureSimilar attack · NVD/CVE Database
- HighCVE-2026-118625: MLflow Tracking Server privilege escalation via get_endpointsSimilar attack · NVD/CVE Database
- InfoWhen trust becomes the payload: Inside the threats, etc quarterly reportSimilar attack · Elastic Security Labs
- MediumCritical Langflow flaws enable authentication bypass and RCESimilar attack · CSO Online