Executive brief
TypeBot is an open-source chatbot builder used to create conversational interfaces. A security flaw in the bot engine allows an attacker to potentially access chat results and session data belonging to other users. While the risk is lowered by the difficulty of guessing specific session IDs, a successful exploit could expose sensitive information such as names, emails, and phone numbers submitted during previous chat sessions.
Technical details
A missing authorization check in the `findResult` Prisma query allows for cross-tenant data access. The query retrieves chat results based solely on a `resultId` without verifying that the result belongs to the specific `typebotId` being accessed. An attacker can exploit this by providing a foreign `resultId` to the `startChat` endpoint. Successful exploitation requires the target chatbot to have the `rememberUser` setting enabled and requires the attacker to know or guess a valid 24-character CUID2 identifier. If these conditions are met, the attacker can retrieve previous user answers, session variables, and PII. The issue is fixed in version 3.16.0 by enforcing a `typebotId` filter in the database query.
Affected products
- baptisteArno TypeBot <= 3.15.2
Timeline
- 2026-04-08: patched: Version 3.16.0 released
- 2026-05-22: disclosed: Security advisory published
- 2026-05-22: advisory: CVE-2026-39967 assigned