Executive brief
AgentChat, a communication platform, contains a security flaw in its user information service. An unauthorized person can access sensitive account details, including password hashes, for any user on the system. This could allow an attacker to take over user accounts and access private conversations or data.
Technical details
An information disclosure vulnerability exists in AgentChat versions up to and including v2.3.0 due to improper data filtering in the /api/v1/user/info endpoint. The backend implementation in user.py and UserService fails to sanitize the database object before returning it to the client, resulting in the exposure of all UserTable fields. An unauthenticated attacker can perform user ID enumeration to collect SHA256 password hashes. Because these hashes are unsalted, they are highly susceptible to rainbow table or brute-force attacks. Recommended mitigations include implementing response models that exclude sensitive fields and migrating to salted hashing algorithms like bcrypt.
Affected products
- AgentChat AgentChat <= v2.3.0
Timeline
- 2026-06-09: disclosed: CVE published to NVD