Intel

AIKIDO-2026-459021

google-adk is vulnerable to Server-Side Request Forgery (SSRF)

Server-Side Request Forgery (SSRF) Pre-CVE
Found by Aikido Intel before public disclosure or CVE publication.
Published Aug 10, 2026

32

Low Risk

This Affects:

PYTHONgoogle-adk
0.0.2 - 2.5.0
Fixed in 2.6.0
Are you affected? Scan for Free

TL;DR

The LlmAgent accepts a generate_content_config whose http_options.base_url is copied into every model request and overrides the client transport. When an agent configuration is derived from untrusted input, such as a loaded agent YAML config, this redirects a credentialed model call to a host of the config author's choosing, sending the configured API key and the full prompt and response traffic to that endpoint. The fix rejects base_url in generate_content_config while leaving request-time headers, timeout, and retry options unaffected.

Who does this affect?

You are affected if you are using a version that falls within the vulnerable range and your application builds an agent from a generate_content_config that is derived from untrusted or externally supplied input, such as a loaded agent YAML configuration.

Background info

google-adk is vulnerable to Server-Side Request Forgery (SSRF) in versions 0.0.2 - 2.5.0.

How to fix this

Upgrade the google-adk library to the patch version.