MediumVulnerability
CVE-2026-54249: Pydantic AI UI adapters forward unvalidated UploadedFile references
- Source
- NVD(opens in a new tab)
- Identifier
- CVE-2026-54249
- Published
- Record updated
- Affected
- pydantic-ai-slim >= 1.65.0, < 1.106.0, fixed in 1.106.0, released June 4, 2026, 54 days before disclosure
- pydantic-ai-slim >= 2.0.0b1, < 2.0.0b6, fixed in 2.0.0b6, released June 4, 2026, 54 days before disclosure
- pydantic-ai >= 1.65.0, < 1.106.0, fixed in 1.106.0, released June 4, 2026, 54 days before disclosure
- and 1 more
- Known exploitation
- Not listed in the CISA Known Exploited Vulnerabilities catalog at the last check.
- EPSS
- 0.3%
Summary
Pydantic AI versions 1.65.0 through 1.105.0, and 2.0.0b1 through 2.0.0b5, do not validate UploadedFile references in message history submitted to UI adapters such as the Vercel AI adapter. A client can supply a provider file ID or cloud-storage URI (for example s3:// or gs://), and the provider resolves it with the server's own identity, letting an attacker read objects the server can access, including other tenants' objects. Exploitation requires a valid file identifier, which may be guessable depending on how the application names objects.
Mitigation
Fixed in 1.106.0 and 2.0.0b6.
Affected packages in the Exposure Registry
Matched by package name and ecosystem. Each entry shows whether the package delegates to a language model and how many tracked packages depend on it.
- pydantic-aiPyPILLM dependency since 2024-05-20 · 4 tracked dependents
- pydantic-ai-slimPyPILLM dependency since 2024-11-25 · 7 tracked dependents
Topics
Related items
- MediumCVE-2026-121936: easyVoice server-side request forgery through openaiBaseUrl in /api/v1/ttsSimilar attack · NVD/CVE Database
- MediumCVE-2026-118628: MLflow get-artifacts server-side request forgery information disclosureSimilar attack · NVD/CVE Database
- HighCVE-2026-118625: MLflow Tracking Server privilege escalation via get_endpointsSimilar attack · NVD/CVE Database
- InfoWhen trust becomes the payload: Inside the threats, etc quarterly reportSimilar attack · Elastic Security Labs
- MediumCritical Langflow flaws enable authentication bypass and RCESimilar attack · CSO Online