Intel

AIKIDO-2026-357199

pydantic-ai-backend is vulnerable to Improper Access Control

Improper Access Control Pre-CVE
Found by Aikido Intel before public disclosure or CVE publication.
Published 2 days ago

59

Medium Risk

This Affects:

PYTHONpydantic-ai-backend
0.1.2 - 0.2.15
Fixed in 0.2.16
Are you affected? Scan for Free

TL;DR

The LocalBackend permission ruleset only consulted deny rules on the read, write, and edit operations and the execute command-pattern check, while content-returning paths ignored the ruleset. read_bytes, grep_raw, and ls_info/glob_info returned file bytes, search matches, and directory listings for paths a configured deny rule was meant to block, so a rule such as **/restricted/** or **/.env* could be bypassed to read, search, or list restricted files. The fix applies the same read/grep/ls/glob deny checks on every content-returning path and adds a best-effort path guard on the execute paths. Denied reads now return empty bytes, denied grep matches are hidden, and denied listing entries are omitted.

Who does this affect?

You are affected if you are using a version that falls within the vulnerable range and have configured a permission ruleset with deny rules on a LocalBackend.

Background info

pydantic-ai-backend is vulnerable to Improper Access Control in versions 0.1.2 - 0.2.15.

How to fix this

Upgrade the pydantic-ai-backend library to the patch version.