Description

A vulnerability has been found in gomarble-ai facebook-ads-mcp-server 0.1.0. The impacted element is the function fetch_pagination_url of the file server.py. Such manipulation leads to server-side request forgery. The attack can be launched remotely. The name of the patch is 4e53875aa22e8991c2fa4a7660d86e1caba66659. Applying a patch is advised to resolve this issue.

Executive Summary

Technical Details

Mitigations

References

Additional Metadata

{
  "score": "6.3",
  "pubDate": "2026-08-16T21:16:37.913Z",
  "pubdate": "2026-08-16T21:16:37.913Z",
  "executiveSummary": "A Server-Side Request Forgery (SSRF) vulnerability has been identified in the gomarble-ai facebook-ads-mcp-server version 0.1.0.\nThe flaw resides within the fetch_pagination_url function located in server.py, allowing remote attackers to manipulate input parameters and induce the server into initiating unauthorized outbound HTTP requests.\nThis vulnerability poses a significant risk to internal network security, as successful exploitation may enable attackers to probe internal network services, access metadata endpoints, or interact with external systems via the vulnerable server.\nThe attack can be executed remotely without requiring prior authentication or privileged access, relying solely on the application's improper handling of user-supplied pagination URLs.\nMitigation requires immediate remediation through the application of the official security patch identified by commit hash 4e53875aa22e8991c2fa4a7660d86e1caba66659.",
  "technicalDetails": "The vulnerability is classified as Server-Side Request Forgery (SSRF), stemming from inadequate validation and sanitization of user-supplied input utilized in outbound request construction.\nThe vulnerable component is specifically the fetch_pagination_url function implemented within the server.py file of the gomarble-ai facebook-ads-mcp-server version 0.1.0 codebase.\nThe root cause lies in the application's logic directly consuming pagination URLs provided in API responses or input parameters without enforcing strict URL parsing, scheme restrictions, or destination whitelisting prior to execution.\nAttack flow begins when a remote, unauthenticated attacker interacts with the service and supplies a maliciously crafted pagination URL designed to target internal network resources, loopback interfaces (e.g., 127.0.0.1 or ::1), or cloud metadata services.\nUpon receiving the malicious input, the fetch_pagination_url function processes the URL and dispatches an outbound HTTP request from the host executing the server.py script.\nBecause the network exposure of the service allows remote interaction, the attacker can leverage the server as a proxy to bypass perimeter network defenses and access restricted internal assets that are otherwise unreachable from the public internet.\nPost-exploitation impact depends on the internal network topology, potentially leading to unauthorized data exfiltration, internal service enumeration, or interaction with sensitive internal application programming interfaces."
}