CVEvulnerability

CVE-2026-107288

3.7 LOW Published 2026-10-08 Modified 2026-10-13 Official page ↗
Status
Analyzed
Published
2026-10-08
Last modified
2026-10-13
Assigner
[email protected]
CVSS 3.1 ([email protected])
3.7 LOW — CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:N/A:N
Weaknesses
CWE-1289, CWE-918

Description

Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.77.0 until 1.107.6 and 2.44.0, the local web_fetch_tool and the WebFetch local fallback compare blocked_domains entries with a URL hostname before both values are normalized to the form used by getaddrinfo. An attacker-influenced model can use an equivalent IDNA spelling, non-ASCII label separator, case variation, or trailing root label that resolves to a blocked host but does not match the configured string, causing the application to fetch that host with its own privileges. allowed_domains fails closed for unmatched spellings, and private-IP and cloud-metadata protections remain effective. This issue is fixed in versions 1.107.6 and 2.44.0.

Affected configurations

• pydantic pydantic ai (>= 1.77.0 < 1.107.6) • pydantic pydantic ai (>= 2.0.0 < 2.44.0)

Related

Weakness · CWE (2)

  • CWE-1289 Improper Validation of Unsafe Equivalence in Input weakness
  • CWE-918 Server-Side Request Forgery weakness

References