Intel

AIKIDO-2026-11012

pydantic-deep is vulnerable to Protection Mechanism Failure

Protection Mechanism Failure Pre-CVE
Found by Aikido Intel before public disclosure or CVE publication.
Published Yesterday

51

Medium Risk

This Affects:

PYTHONpydantic-deep
0.1.0 - 0.3.22
Fixed in 0.3.23
Are you affected? Scan for Free

TL;DR

The browser allowlist was only enforced on direct navigate() calls, leaving other navigation paths such as link clicks, JavaScript redirects, and history navigation outside the restriction boundary. As a result, a model or attacker controlling browser actions could still reach disallowed domains and interact with unintended external content despite an allowed_domains policy being configured. An attacker might exploit this by placing a cross-domain link on an allowed page, triggering execute_js to set location.href to an attacker-controlled site, or using go_back / go_forward to load previously visited disallowed pages, potentially exposing the model to malicious content, data exfiltration workflows, or policy bypass. The fix adds defense-in-depth by blocking top-level disallowed navigations at the network layer and validating the final page URL after each navigation-capable action.

Who does this affect?

You are affected if you are using a version that falls within the vulnerable range and rely on allowed_domains for the browser capability, any user/model-driven navigation via click, execute_js, history, or popups could bypass the boundary.

Background info

pydantic-deep is vulnerable to Protection Mechanism Failure in versions 0.1.0 - 0.3.22.

How to fix this

Upgrade the pydantic-deep library to the patch version.