TransformerOptimus SuperAGI v0.0.14 contains an improper access control flaw in the delete_user_knowledge handler that allows any client to delete knowledge entries without authenticating or verifying that the requester owns the organization associated with the knowledge. This flaw can lead to loss or corruption of data that is central to the AI model’s operation, potentially disrupting downstream services that depend on the knowledge base.

The vulnerability is present in the SuperAGI v0.0.14 source, specifically within superagi/controllers/knowledges.py. Systems running this exact version or earlier source snapshots that have not been patched with a release enforcing proper authentication and ownership checks are affected.

An attacker only needs network access to the host exposing the API; a simple HTTP POST to /knowledges/delete/{knowledge_id} will trigger deletion because no authentication is required and no ownership validation is performed. The CVSS score of 7.5 indicates high severity. No EPSS data is available, and the vulnerability is not listed in CISA KEV, yet the lack of authentication coupled with the endpoint’s public exposure means the risk of exploitation is significant.