Description

Authorization bypass in the process and annotation taxonomy serializers in Scripta eScriptorium through 26.04.1 allows a remote authenticated user to run segmentation and transcription against other users' document parts, overwriting their content, via part primary keys supplied to a many=True related field whose queryset restriction was applied to the ManyRelatedField instead of its child_relation and therefore had no effect

Executive Summary

Technical Details

Mitigations

References

Additional Metadata

{
  "score": "6.5",
  "pubDate": "2026-08-06T16:16:37.853Z",
  "pubdate": "2026-08-06T16:16:37.853Z",
  "executiveSummary": "An authorization bypass vulnerability exists in the process and annotation taxonomy serializers within Scripta eScriptorium through version 26.04.1.\nThe flaw allows a remote authenticated attacker to execute segmentation and transcription processes against document parts belonging to other users, resulting in the unauthorized overwriting of sensitive document content.\nThe risk implication is significant as it compromises data integrity and confidentiality across multi-tenant environments by permitting cross-user resource manipulation.\nExploitation requires the attacker to be authenticated within the application and to supply arbitrary primary keys corresponding to target document parts.\nThe vulnerability stems from improper configuration of related field querysets within the Django REST Framework serializers, where restrictions were incorrectly applied to the parent container rather than the child relation.",
  "technicalDetails": "The vulnerability resides in the process and annotation taxonomy serializers of Scripta eScriptorium through version 26.04.1, specifically affecting how related fields handle queryset scoping.\nThe root cause of the authorization bypass is an implementation error in Django REST Framework where queryset restrictions were applied to the ManyRelatedField instead of its underlying child_relation. Consequently, the security constraint had no operational effect, failing to filter out unauthorized primary keys.\nThe attack flow proceeds as follows: First, a remote authenticated user initiates an API request targeting the segmentation or transcription processes. Second, the user supplies arbitrary document part primary keys within a many=True related field. Third, because the queryset restriction is misapplied to the ManyRelatedField, the input validation fails to verify whether the supplied primary keys belong to the authenticated user.\nAs a result, the application processes the request against foreign document parts. The payload behavior executes backend segmentation and transcription tasks, ultimately overwriting the victim user's existing document content with newly processed data.\nPrerequisites for exploitation include network exposure of the application API, a valid low-privileged or standard authenticated user session, and knowledge or enumeration of target document part primary keys.\nThe post-exploitation impact includes unauthorized data modification, potential denial of service via data destruction for legitimate users, and severe integrity violations within collaborative or multi-tenant deployments of Scripta eScriptorium."
}