CVE-2026-7669— sglang 注入漏洞
一分钟漏洞结论
影响对象
sgl-project SGLang
利用判断
尚无明确在野利用证据，仍需结合暴露面评估
建议动作
优先检查厂商安全公告和参考链接中的修复版本；无法立即升级时，限制受影响服务暴露并加强监测。
sglang是sgl-project开源的一个用于加速大模型推理的编程语言与运行时系统。 SGLang 0.5.9及之前版本存在注入漏洞，该漏洞源于python/sglang/srt/utils/hf_transformers_utils.py文件中get_tokenizer函数对参数trust_remote_code的操作，可能导致代码注入。
影响版本矩阵 10
| 厂商 | 产品 | 版本范围 | 状态 |
|---|---|---|---|
| sgl-project | SGLang | 0.5.0 |
affected |
0.5.1 |
affected | ||
0.5.2 |
affected | ||
0.5.3 |
affected | ||
0.5.4 |
affected | ||
0.5.5 |
affected | ||
0.5.6 |
affected | ||
0.5.7 |
affected | ||
| … +2 条更多 | |||
获取后续新漏洞提醒
登录后订阅
一、 漏洞 CVE-2026-7669 基础信息
漏洞信息
对漏洞内容有疑问？看看神龙的深度分析是否有帮助！
查看神龙十问 ↗
尽管我们使用了先进的大模型技术，但其输出仍可能包含不准确或过时的信息。神龙努力确保数据的准确性，但请您根据实际情况进行核实和判断。
Vulnerability Title
sgl-project SGLang HuggingFace Transformer hf_transformers_utils.py get_tokenizer code injection
来源: CVE Program / CVE List V5
Vulnerability Description
A vulnerability was detected in sgl-project SGLang up to 0.5.9. Impacted is the function get_tokenizer of the file python/sglang/srt/utils/hf_transformers_utils.py of the component HuggingFace Transformer Handler. The manipulation of the argument trust_remote_code with the input False as part of Boolean results in code injection. The attack can be executed remotely. A high complexity level is associated with this attack. The exploitability is considered difficult. In get_tokenizer(), when the caller passes trust_remote_code=False and HuggingFace transformers v5 returns a TokenizersBackend instance (the generic fallback for tokenizer classes not in the registry), SGLang silently re-invokes AutoTokenizer.from_pretrained with trust_remote_code=True, overriding the caller's explicit security setting. A model repository containing a malicious tokenizer.py referenced via auto_map in tokenizer_config.json will execute arbitrary Python in the SGLang process during this second call. No log line or warning is emitted. The override affects all current SGLang versions because transformers==5.3.0 is pinned in pyproject.toml. Both tokenizer_mode="auto" and tokenizer_mode="slow" are affected. The exploit is now public and may be used. The vendor was contacted early about this disclosure but did not respond in any way.
来源: CVE Program / CVE List V5
CVSS Information
CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:L/A:L
来源: CVE Program / CVE List V5
Vulnerability Type
对生成代码的控制不恰当(代码注入)
来源: CVE Program / CVE List V5
Vulnerability Title
sglang 注入漏洞
来源: 中国国家信息安全漏洞库 CNNVD
Vulnerability Description
sglang是sgl-project开源的一个用于加速大模型推理的编程语言与运行时系统。 SGLang 0.5.9及之前版本存在注入漏洞，该漏洞源于python/sglang/srt/utils/hf_transformers_utils.py文件中get_tokenizer函数对参数trust_remote_code的操作，可能导致代码注入。
来源: 中国国家信息安全漏洞库 CNNVD
CVSS Information
N/A
来源: 中国国家信息安全漏洞库 CNNVD
Vulnerability Type
N/A
来源: 中国国家信息安全漏洞库 CNNVD
受影响产品
| 厂商 | 产品 | 影响版本 | CPE | 订阅 |
|---|---|---|---|---|
| sgl-project | SGLang | 0.5.0 | - |
|
二、漏洞 CVE-2026-7669 的公开POC
| # | POC 描述 | 源链接 | 神龙链接 |
|---|
AI 生成 POC
高级
未找到公开 POC。
登录以生成 AI POC三、漏洞 CVE-2026-7669 的情报信息
请登录查看更多情报信息。
CVE-2026-7669 公开利用代码 (1)
IV. Related Vulnerabilities
Same vendor: sgl-project
V. Comments for CVE-2026-7669
暂无评论