Executive brief
Flowise, a document-store and chatflow orchestration platform, exposes APIs that retrieve and delete document processing history without proper permission checks. An authenticated user or API key from one workspace can read or delete another workspace's document upsert history, including sensitive configuration data for embeddings and vector stores, if they know the target chatflow ID—which is publicly disclosed in shareable chatbot links.
Technical details
The GET /api/v1/upsert-history/:id and PATCH /api/v1/upsert-history endpoints lack route-level permission checks and the backing service performs no workspace or ownership validation. An attacker with any valid authenticated session or API key can query UpsertHistory rows by chatflowId and delete rows by UUID, since these operations are scoped only to attacker-supplied values. The exposed flowData and result fields leak embedding, record-manager, and vector-store node configurations including credential paramValues.
Affected products
- Flowise AI Flowise through 3.1.4
Timeline
- 2026-09-26: disclosed
- 2026-09-10: advisory