Authorization Bypass in FastGPT by LabRing
CVE-2026-61643

5.9MEDIUM

Key Information:

Vendor

Labring

Status
Fastgpt
Vendor
CVE Published:
15 July 2026

What is CVE-2026-61643?

A significant vulnerability exists in the FastGPT platform, affecting versions 4.14.17 through 4.15.0-beta5. This issue arises when an authenticated user can create a workflow that references another user's private HTTP toolset. By leveraging a specially crafted tool ID, an attacker can bypass standard access controls, permitting unauthorized execution of private tools. The flaw allows an attacker to successfully relay requests to /api/v2/chat/completions, executing the victim's private HTTP tool. This issue has been addressed in version 4.15.0-beta5, which includes enhanced authorization checks during workflow saves.

Affected Version(s)

FastGPT >= 4.14.17, < 4.15.0-beta5

References

https://github.com/labring/FastGPT/security/advisories/GH...
x_refsource_CONFIRM
https://github.com/labring/FastGPT/releases/tag/v4.15.0-b...
x_refsource_MISC

CVSS V3.1

Score:
5.9
Severity:
MEDIUM
Confidentiality:
Low
Integrity:
High
Availability:
Low
Attack Vector:
Network
Attack Complexity:
High
Privileges Required:
Low
User Interaction:
None
Scope:
Unchanged

Timeline

  • Vulnerability published

  • Vulnerability Reserved

.