CVE-2026-119048 in JeeSite
Summary
by MITRE • 10/29/2026
JeeSite 5.11.0 through 5.19.1 contains an insecure direct object reference vulnerability in the AI CMS chat module that allows authenticated users to access other users' conversations due to missing and reversed ownership checks. Attackers who know or derive a victim's conversation id can call /a/cms/chat/message, /a/cms/chat/delete, or /a/cms/chat/completions to read, delete, or append to the victim's AI chat transcripts.
If you want to get best quality of vulnerability data, you may have to visit VulDB.
Analysis
by VulDB Data Team • 10/29/2026
The vulnerability identified in JeeSite versions 5.11.0 through 5.19.1 represents a critical failure in access control mechanisms within its integrated Artificial Intelligence CMS chat module. This flaw is classified as an Insecure Direct Object Reference, commonly referred to by the acronym IDOR or CWE-639. The core technical issue stems from insufficient validation of object ownership during API requests. Specifically, when authenticated users interact with endpoints such as /a/cms/chat/message, /a/cms/chat/delete, and /a/cms/chat/completions, the application fails to verify that the requesting user is the legitimate owner of the conversation ID being accessed or modified. Instead of enforcing strict server-side authorization checks that bind the session identity to the resource ownership, the system relies on client-supplied identifiers without adequate cross-referencing against a secure database record of permissions. This architectural oversight allows any authenticated actor who can guess or derive another user's unique conversation identifier to bypass intended privacy boundaries.
The operational impact of this vulnerability is severe due to the sensitive nature of AI chat transcripts, which often contain proprietary business information, personal data, or confidential strategic discussions. An attacker with valid credentials for a low-privilege account can exploit this flaw to perform unauthorized read operations on other users' conversations by invoking the message retrieval endpoint. Furthermore, the lack of ownership verification extends to destructive and manipulative actions; attackers can delete existing conversation history using the delete endpoint, potentially causing data loss or hindering forensic analysis in security incidents. Additionally, the ability to append messages via the completions endpoint poses a significant risk for prompt injection attacks or social engineering within the organization's internal communication channels. By injecting malicious prompts into another user's chat context, an attacker could manipulate the AI assistant to generate harmful content, leak sensitive information from its training data if exposed through the victim's session, or disrupt normal workflow operations.
From a threat modeling perspective, this vulnerability aligns with MITRE ATT&CK technique T1078, specifically Valid Accounts and potentially T1530 Data from Local System Deserialization if combined with other exploits, though primarily it falls under unauthorized data access and modification. The exploitation requires the attacker to possess valid authentication credentials for any user account within the JeeSite environment, which lowers the barrier to entry compared to unauthenticated attacks but remains highly feasible in environments where credential sharing or phishing is common. Once authenticated, the attack does not require complex privilege escalation steps beyond deriving the target conversation ID, making it a straightforward and high-impact exploitation path for insider threats or compromised accounts.
Mitigation strategies must focus on implementing robust server-side authorization checks at every API endpoint handling sensitive resources. Developers should ensure that each request to chat-related endpoints includes validation logic that queries the database to confirm the authenticated user's identity matches the owner field of the requested conversation record before processing any read, write, or delete operations. This principle of least privilege and explicit ownership verification is essential for preventing IDOR vulnerabilities. Additionally, implementing rate limiting on these endpoints can help mitigate automated enumeration attacks used to discover valid conversation IDs. For existing deployments within the affected version range, immediate patching to a version beyond 5.19.1 where this issue has been resolved by the vendor is the most effective remediation step. Until patches are applied, network-level controls such as Web Application Firewalls may offer limited protection if they can detect anomalous patterns of ID enumeration or unauthorized access attempts based on known attack signatures, though application-layer fixes remain mandatory for complete resolution.