Improper Access Control in auth0-ai-python Document Retrieval CVE-2026-108167 - Nov 10, 2026

Description

The auth0-ai-python FGA retriever does not distinguish between authorization checks under different users, relations, or conditions when filtering documents during retrieval-augmented generation, resulting in documents denied by Auth0 FGA being included in the model context and used in the generated response.

Preconditions

1. An application uses the auth0-ai-langchain or auth0-ai-llamaindex FGA retriever to filter documents, and

2. A single retrieval includes multiple documents that map to the same Auth0 FGA object, but are associated with different users, relations, or authorization contexts.

Affected product and versions

Customers using auth0-ai-langchain prior to 1.0.2, or auth0-ai-llamaindex versions prior to 1.0.3 are affected.

Customer Recommendations

To remediate this vulnerability, upgrade auth0-ai-langchain to version 1.0.2 or greater and auth0-ai-llamaindex to version 1.0.3 or greater.

Resolution

The vulnerability is present in auth0-ai-langchain versions prior to 1.0.1 and auth0-ai-llamaindex versions prior to 1.0.2 and is resolved in auth0-ai-langchain version 1.0.2 and auth0-ai-llamaindex version 1.0.3.

CVE details

References

Improper Access Control in auth0-ai-python Document Retrieval