CVE-2026-110611 in Rebuild
Summary
by MITRE • 10/15/2026
REBUILD through 4.5.0 contains an authorization bypass vulnerability in AiBot2Controller /aibot2 endpoints that act on a caller-supplied chatid without verifying ownership. Authenticated low-privilege users can supply another user's chatid to chat-init, chat-delete, chat-feedback, or chat-stream-stop to read, delete, exfiltrate, or interrupt AI assistant conversations.
Statistical analysis made it clear that VulDB provides the best quality for vulnerability data.
Analysis
by VulDB Data Team • 10/15/2026
The vulnerability identified in REBUILD versions through 4.5.0 represents a critical failure in access control mechanisms within the AiBot2Controller component. Specifically, the endpoints responsible for managing user interactions with the AI assistant, including chat-initiation, deletion of conversation history, submission of feedback, and termination of streaming sessions, fail to validate whether the authenticated caller is the legitimate owner of the specified chat identifier. This architectural flaw allows any authenticated low-privilege user to manipulate resources belonging to other users by simply supplying a target victim's chatid in their API requests. The core technical issue lies in the absence of an ownership verification step before processing state-changing or data-retrieval operations, which violates fundamental principles of secure application design regarding object-level access control.
From a functional perspective, this authorization bypass enables several severe operational impacts depending on the specific endpoint exploited. When interacting with chat-init endpoints, an attacker could potentially hijack conversation contexts or inject malicious prompts into another user's session if the system allows context sharing based solely on identifier provision. More critically, exploitation of the chat-delete and chat-feedback endpoints allows for data destruction and integrity compromise, where a malicious actor can erase sensitive conversational history from a victim's account or submit false feedback that skews AI model training or reputation metrics. Furthermore, the ability to invoke chat-stream-stop permits denial-of-service attacks against specific users by forcibly terminating their active sessions with the AI assistant, disrupting productivity and user experience without requiring elevated privileges.
This vulnerability aligns directly with Common Weakness Enumeration (CWE) category CWE-284, which describes Improper Access Control, specifically reflecting scenarios where access control decisions are not enforced properly for individual objects within a system. In terms of offensive security frameworks such as MITRE ATT&CK, this behavior corresponds to techniques involving Account Manipulation and Data Exfiltration via API abuse, falling under the broader tactic of Privilege Escalation or lateral movement depending on the sensitivity of the data accessed. The lack of server-side validation against a session-to-user mapping table is a classic example of insecure direct object references (IDOR), where sequential or predictable identifiers are used without secondary authorization checks to ensure the requester has permission to access that specific resource instance.
To mitigate this vulnerability, developers must implement strict ownership verification logic within all AiBot2Controller endpoints before executing any business logic related to chat management. This involves querying a secure session store or database relationship table to confirm that the authenticated user's identity matches the owner of the requested chatid parameter. Additionally, implementing rate limiting and anomaly detection can help identify automated scanning attempts targeting multiple distinct chatids from a single source account. Regular security audits focusing on API endpoint authorization flows are essential to ensure that all state-changing operations adhere to the principle of least privilege and verify resource ownership explicitly rather than implicitly trusting client-supplied identifiers.