CVE-2026-54249

VercelAIAdapter trusts client-controlled `providerMetadata` to construct `UploadedFile` — S3/GCS confused deputy via provider metadata injection

MEDIUM
6.8
CVSS Score
Published: July 29, 2026 at 08:28 PM UTC
Updated: August 4, 2026 at 10:03 PM UTC
1 product affected across 1 vendorPydantic AI

Using pydantic-ai by pydantic?

Get instant alerts when new vulnerabilities affect pydantic-ai - before attackers find them.

Real-time alertsTrack your software stackFree plan available
Severity
MEDIUM6.8
Exploit prediction
EPSS 0.3%
SSVC
No known exploit

Description

Pydantic AI is a Python agent framework for building Generative AI applications. In versions 1.65.0 through 1.105.0, and 2.0.0b1 through 2.0.0b5, a client that submits message history to a Pydantic AI UI adapter (such as the Vercel AI adapter) can reference arbitrary files in the application's model-provider or cloud-storage account. While file URL parts are validated against a scheme allowlist, UploadedFile references — which point to a file by provider file ID or cloud-storage URI (e.g. s3://…, gs://…) — were forwarded without validation. Because the provider resolves an UploadedFile using the server-side identity (IAM role, service account, or provider API key) rather than the client's, an attacker can craft message history to make the server read objects from its own account or other tenants, given a referenceable identifier. Exploitation requires a valid file identifier, which is not always unguessable depending on how the application names objects. This issue has been fixed in versions 1.106.0 and 2.0.0b6.

Affected

Our own answer: ranges merged, affected separated from platform-only, with confidence and review flags.

v < 1.106.0v < 2.0.0b61.65.0 ≤ v < 1.105.02.0.0 beta12.0.0 beta22.0.0 beta32.0.0 beta42.0.0 beta5
Affected range stops below 1.105.0, so it looks safe — our inference, not a vendor statement
Affected range stops below 1.106.0, so it looks safe — our inference, not a vendor statement
Affected range stops below 2.0.0b6, so it looks safe — our inference, not a vendor statement

Remediation

No guide yet

No remediation guidance has been written for this CVE yet. What the record does show:

References (1)

SSVC — Action Prioritization

SSVCNo known exploitNot automatablePartial impact

SSVC signals action urgency (what to patch first), separate from CVSS which measures severity.

CVSS

6.8MEDIUMGitHub_M
Attack Vector
NETWORK
Complexity
HIGH
Privileges
NONE
User Interaction
NONE
Scope
CHANGED
Confidentiality
HIGH
Integrity
NONE
Availability
NONE
Vector?

Quick Info

StatusANALYZED
CWE
CWE-918

Get alerted

Stop reading CVE feeds looking for your own stack

Track the products you run and VulnTracker tells you when a new CVE affects them — with what is known about fixing it, and nothing about the rest.

Free plan, no card