Vulnerability in Pydantic AI Framework Affecting Generative AI Applications
CVE-2026-65975

6.5MEDIUM

Key Information:

Vendor

Pydantic

Status
Pydantic-ai
Pydantic-ai-slim
Vendor
CVE Published:
29 July 2026

What is CVE-2026-65975?

A vulnerability in the Pydantic AI framework allows remote clients to execute non-approved tools with unvalidated client-supplied arguments due to insufficient sanitization of unresolved tool calls. This security flaw arises from an improper handling of message sanitization, which fails to effectively prevent the dispatch of potentially harmful commands. As a result, the integrity and security of applications built on this framework could be severely compromised, emphasizing the importance of updating to patched versions 1.107.1 and 2.5.0 to mitigate potential risks.

Affected Version(s)

pydantic-ai >= 1.88.0, < 1.107.1 < 1.88.0, 1.107.1

pydantic-ai >= 2.0.0b1, < 2.5.0 < 2.0.0b1, 2.5.0

pydantic-ai-slim >= 1.88.0, < 1.107.1 < 1.88.0, 1.107.1

References

https://github.com/pydantic/pydantic-ai/security/advisori...
x_refsource_CONFIRM

CVSS V3.1

Score:
6.5
Severity:
MEDIUM
Confidentiality:
Low
Integrity:
Low
Availability:
Low
Attack Vector:
Network
Attack Complexity:
Low
Privileges Required:
None
User Interaction:
None
Scope:
Unchanged

Timeline

  • Vulnerability published

  • Vulnerability Reserved

.