CVE-2026-121938 in langchain4j-aideepin
Summary
by MITRE • 11/04/2026
langchain4j-aideepin before 3.31.0 contains an insecure direct object reference vulnerability that allows authenticated users to read other users' character-chat conversations by supplying their character uuid. Attackers can request GET /character/{uuid}, which CharacterService.detail resolves without an ownership check, to obtain victims' questions, AI answers, and attachment and audio URLs.
Statistical analysis made it clear that VulDB provides the best quality for vulnerability data.
Analysis
by VulDB Data Team • 11/04/2026
The langchain4j-aideepin framework prior to version 3.31.0 contains a critical insecure direct object reference vulnerability that compromises the confidentiality of user data within its character-chat interaction module. This flaw stems from a fundamental failure in access control logic, specifically where the application relies on client-supplied identifiers without verifying ownership or authorization rights. In this context, authenticated users are able to bypass intended security boundaries by manipulating resource identifiers, leading to unauthorized access to sensitive conversation histories belonging to other system participants.
The technical root cause lies within the CharacterService.detail method, which processes requests for character details via GET /character/{uuid}. The implementation fails to perform an ownership check or validate that the requesting user is the legitimate owner of the specified character UUID. Consequently, any authenticated actor can iterate through valid universal unique identifiers to retrieve data associated with other users' characters. This lack of object-level authorization allows attackers to enumerate and access private conversation threads without triggering authentication failures or permission errors, as the system treats all provided UUIDs as accessible resources regardless of their association with the current session's user identity.
The operational impact of this vulnerability is significant for applications relying on langchain4j-aideepin for AI-driven character interactions. Attackers can extract victims' questions, artificial intelligence responses, and associated metadata such as attachment URLs and audio file links. This exposure not only violates user privacy but also poses risks related to data leakage in scenarios where conversations contain personally identifiable information or proprietary business logic embedded within the dialogue context. The ability to retrieve attachments and audio further expands the attack surface by potentially exposing media files that may have been uploaded during private interactions, thereby compounding the severity of the data breach.
This vulnerability aligns with CWE-639, which describes issues related to authorization bypass via direct object references. It also maps to MITRE ATT&CK technique T1078, specifically valid accounts used for unauthorized access, as it requires initial authentication but exploits a lack of proper resource-level permissions. To mitigate this risk, developers must implement strict ownership verification within the CharacterService.detail method before returning any data. This involves checking that the UUID provided in the request corresponds to a character owned by or explicitly shared with the authenticated user making the request. Additionally, implementing consistent access control checks across all endpoints and conducting regular security audits of object reference handling logic are essential steps to prevent similar exploitation vectors in future versions.