CVE-2026-65975: pydantic ai Auth Bypass — Fix

Description

Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. In versions 1.88.0 up to but not including 1.107.1 and 2.0.0b1 up to but not including 2.5.0, the UI adapters (AG-UI via Agent.to_ag_ui()/AGUIAdapter, and Vercel AI via VercelAIAdapter) use sanitize_messages to strip unresolved ("dangling") client-submitted tool calls from untrusted message history before it reaches the agent, a defense-in-depth default that prevents the agent from executing tool calls the model never emitted. However, the strip anchored to a message index computed before sanitization ran, so when a trailing client message sanitized to empty and was dropped (for example a client system message under the default manage_system_prompt='server'), a preceding assistant response carrying an unresolved tool call became the new tail and was dispatched without inspection. As a result, a remote client could cause a registered, non-approval server tool to run with client-supplied arguments rather than arguments the model produced. The impact is bounded by what the affected tools do and is most significant for applications that gate tool execution in a model-request hook (before_model_request / after_model_request), since a forged call skips the model turn and bypasses that guardrail; approval-gated tools (requires_approval=True) are not auto-executed by this path. This issue has been fixed in versions 1.107.1 and 2.5.0.

Is your site exposed to CVE-2026-65975?

Run a free security scan — no signup, results in seconds.

CVSS v3.1 Score

6.5

MEDIUM

CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:N

EPSS — Exploit Prediction

0.0020

Probability of exploitation

0.10%

Percentile rank

EPSS estimates the probability that this vulnerability will be exploited in the wild within the next 30 days. A higher score means more likely to be exploited.

Weakness Type (CWE)

CWE-863 Incorrect Authorization

Affected Products

Vendor	Product	Versions
pydantic	pydantic_ai	1.88.0 — 1.107.1
pydantic	pydantic_ai	2.0.0 — 2.5.0

References

Advisories & Patches

https://github.com/pydantic/pydantic-ai/security/advisories/GHSA-jpr8-2v3g-wgf9

Frequently Asked Questions

What is CVE-2026-65975? +

Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. In versions 1.88.0 up to but not including 1.107.1 and 2.0.0b1 up to but not including 2.5.0, the UI adapters (AG-UI via Agent.to_ag_ui()/AGUIAdapter, and Vercel AI via VercelAIAdapter) use sanitize_messages to strip unresolved ("dangling") client-submitted tool calls from untrusted message history before it reaches the agent, a defense-in-depth default that prevents the agent from executing tool calls the model never emitted. However, the strip anchored to a message index computed before sanitization ran, so when a trailing client message sanitized to empty and was dropped (for example a client system message under the default manage_system_prompt='server'), a preceding assistant response carrying an unresolved tool call became the new tail and was dispatched without inspection. As a result, a remote client could cause a registered, non-approval server tool to run with client-supplied arguments rather than arguments the model produced. The impact is bounded by what the affected tools do and is most significant for applications that gate tool execution in a model-request hook (before_model_request / after_model_request), since a forged call skips the model turn and bypasses that guardrail; approval-gated tools (requires_approval=True) are not auto-executed by this path. This issue has been fixed in versions 1.107.1 and 2.5.0. It has a CVSS v3.1 base score of 6.5 (MEDIUM).

How severe is CVE-2026-65975? +

CVE-2026-65975 has a CVSS v3.1 score of 6.5 out of 10, rated MEDIUM. This is a medium-severity vulnerability that should be remediated as part of regular maintenance. The EPSS score is 0.0020, placing it in the 0th percentile for exploitation probability.

What products are affected by CVE-2026-65975? +

CVE-2026-65975 affects products from pydantic, specifically: pydantic_ai. Check the affected products table above for specific version ranges.

How do I check if I'm vulnerable to CVE-2026-65975? +

You can use Secably's free Website Scanner to check your website for known vulnerabilities. For infrastructure scanning, use the Port Scanner to identify exposed services that may be affected. Check the vendor advisories linked above for specific patch and version information.

Related Vulnerabilities

CVE-2026-65601

Traefik versions 3.7.0 through 3.7.6 contain a namespace confusion vulnerability in the Kubernetes Gateway API provider. When resolving HTTPRoute.spec.rules[].backendRefs[].filters[].extensionRef, Traefik …

CVE-2026-50008

Parse Server is an open source backend that can be deployed to any infrastructure that can run Node.js. From version …

CVE-2026-15829

A SQL injection (CWE-89) and security boundary bypass (CWE-863) vulnerability exists in the prebuilt BigQuery forecasting tool (bigquery-forecast) of googleapis/mcp-toolbox. …

CVE-2026-42571

Pelican is a platform for creating data federations. From versions 7.21.0 to before 7.21.5, 7.22.0 to before 7.22.3, 7.23.0 to …

CVE-2026-42137

Kirby is an open-source content management system. Prior to versions 4.9.0 and 5.4.0, `pages.access/list` and `files.access/list` permissions are not consistently …

CVE-2026-54358

An incorrect authorization vulnerability in MISP allows an organization administrator to target site administrator accounts belonging to the same organization …

CVE-2026-65975