CVE-2025-0740
Published: 30 January 2025
Description
Adversaries may directly collect unsecured credentials stored or passed through user communication services.
Security Summary
CVE-2025-0740, published on 2025-01-30, is an Improper Access Control vulnerability (CWE-284) in EmbedAI versions 2.1 and below. The flaw affects the endpoint "/embedai/chats/load_messages?chat_id=<CHAT_ID>", where inadequate controls on the CHAT_ID parameter enable unauthorized access to sensitive data. It carries a CVSS v3.1 base score of 8.6 (AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:N/A:N), indicating high severity due to network accessibility, low attack complexity, no required privileges or user interaction, and a high confidentiality impact with changed scope.
An authenticated attacker can exploit this vulnerability by simply modifying the CHAT_ID parameter in requests to the affected endpoint, allowing them to retrieve chat messages belonging to other users. This results in unauthorized exposure of potentially sensitive conversation data across user accounts.
The INCIBE-CERT advisory at https://www.incibe.es/en/incibe-cert/notices/aviso/multiple-vulnerabilities-embedai provides further details on this and related vulnerabilities in EmbedAI, including mitigation recommendations.
Details
- CWE(s)
Affected Products
AI Security Analysis
- AI Category
- Enterprise AI Assistants
- Risk Domain
- Privacy and Disclosure
- OWASP Top 10 for LLMs 2025
- None mapped
- MITRE ATLAS Techniques
- None mapped
- Classification Reason
- EmbedAI is a tool for creating chatbots, which aligns with Enterprise AI Assistants as it involves deploying AI-powered conversational interfaces.
MITRE ATT&CK Enterprise Techniques
Why these techniques?
The improper access control (IDOR) in the chat messages endpoint enables authenticated attackers to collect data from other users' messaging/chat sessions (T1213.005) and access potentially unsecured credentials stored in chat messages (T1552.008).