CVE-2026-127222
SGLang through 0.5.23, fixed in commit 7b5fb39, contains a missing authorization vulnerability in the POST /load_lora_adapter_from_tensors HTTP endpoint, which…
Impact
SGLang through 0.5.23, fixed in commit 7b5fb39, contains a missing authorization vulnerability in the POST /load_lora_adapter_from_tensors HTTP endpoint, which lacks the @auth_level(AuthLevel.ADMIN_OPTIONAL) decorator applied to its sibling admin endpoints, allowing unauthenticated network-reachable attackers to invoke it without an API key. Attackers can load a LoRA adapter from attacker-supplied serialized tensors to manipulate the served model, and when combined with a separately documented deserialization bypass, achieve unauthenticated remote code execution as root.
An attacker can exploit this over the network, without logging in, without any action by a user. If it succeeds: high risk of data or systems being changed.
Affected products · Sgl-project
| Product | Affected versions |
|---|---|
| Sglang | through 0.5.23 Version 0 ≤0.5.23 |
Weakness classification
References
- NVD record
- CVE.org record
- GitHub advisory
- https://rahulkarne.com/#/cve/CVE-2026-127222
- https://github.com/rahulreddykarne/CVE-2026-127222-sglang
- https://github.com/sgl-project/sglang/pull/49896
- https://github.com/sgl-project/sglang/commit/7b5fb39bb952da5f6da3bb7fe73c671cf449b6fa
- https://www.vulncheck.com/advisories/sglang-missing-authorization-via-load-lora-adapter-from-tensors
- https://github.com/advisories/GHSA-m386-99gj-pq7p
- https://nvd.nist.gov/vuln/detail/CVE-2026-127222