Executive brief
Typebot, an open-source chatbot builder, contains a security flaw that allows authenticated users to view sensitive data belonging to other accounts. By exploiting a weakness in how the system verifies ownership of chat logs, an attacker could access private information such as AI model outputs, customer data from webhooks, and responses from integrated third-party services (like CRMs or payment processors). This could lead to the exposure of personal information or proprietary business data across different organizations using the platform.
Technical details
An IDOR vulnerability exists in the `getResultLogs` API endpoint within `apps/builder/src/features/results/api/handleGetResultLogs.ts`. While the application correctly verifies that the authenticated user has access to the provided `typebotId`, it fails to validate that the requested `resultId` actually belongs to that specific Typebot. Because the database query for logs is scoped only by `resultId`, an attacker can provide their own authorized `typebotId` alongside a victim's `resultId` to bypass ownership checks. This allows for the retrieval of execution logs which may contain sensitive HTTP response bodies, AI model outputs, and webhook payloads. The issue was addressed by updating the Prisma query to enforce a join between the result and the authorized Typebot ID.
Affected products
- baptisteArno typebot.io <= 3.15.2
Timeline
- 2026-05-22: advisory: GHSA-c63p-mqx5-75r7 published
- 2026-05-22: disclosed
- 2026-04-08: patched: Fixed in version 3.16.0 (referenced as 3.15.2 in some documentation)