Credential Exposure in mcp-from-openapi by AgentFront
CVE-2026-115750

6.3MEDIUM

Key Information:

Vendor

Agentfront

Status
Mcp-from-openapi
Vendor
CVE Published:
23 October 2026

What is CVE-2026-115750?

The mcp-from-openapi library versions 2.5.0 to 2.9.0 are vulnerable to credential exposure due to unsafe handling of Authorization and Cookie headers during cross-origin redirects. Specifically, the function safeFetch() forwards these sensitive headers when redirecting requests, which can be exploited by attackers controlling a spec host or using an open redirect to capture bearer tokens or cookies. It is recommended that users upgrade to version 2.9.0 or later to mitigate this risk.

Affected Version(s)

mcp-from-openapi 2.5.0 < 2.9.0

References

https://github.com/agentfront/mcp-from-openapi/security/a...
vendor-advisory
https://github.com/agentfront/mcp-from-openapi/pull/42
issue-trackingpatch
https://github.com/agentfront/mcp-from-openapi/commit/47f...
patch

CVSS V4

Score:
6.3
Severity:
MEDIUM
Confidentiality:
Low
Integrity:
None
Availability:
None
Attack Vector:
Network
Attack Complexity:
Low
Attack Required:
Physical
Privileges Required:
Undefined
User Interaction:
None

Timeline

  • Vulnerability published

  • Vulnerability Reserved

Credit

Zachariah Mithani (AploSec)
.