Executive brief
Budibase, an open-source low-code platform, contains a security flaw that allows an attacker to hijack a user's account identity. By tricking an authenticated user into clicking a malicious link, an attacker can silently link their own external chat account (such as Slack, Discord, or MS Teams) to the victim's Budibase account. This allows the attacker to impersonate the victim when interacting with Budibase AI agents, potentially gaining unauthorized access to sensitive data or administrative functions.
Technical details
The vulnerability exists in the `GET /api/chat-links/:instance/:token/handoff` endpoint within `packages/server/src/api/controllers/ai/chatIdentityLinks.ts`. This endpoint is public and lacks CSRF protection or a consent UI. An attacker can generate a valid session token via a chat bot's `/link` command and embed their own `externalUserId` in the session. When an authenticated victim visits the crafted URL, the `upsertChatIdentityLink` function is called, mapping the attacker's chat ID to the victim's `globalUserId`. Subsequent interactions from the attacker's chat account are then authorized as the victim. The issue is fixed in version 3.39.0.
Affected products
- Budibase budibase >= 3.37.2, < 3.39.0
Timeline
- 2026-05-28: advisory: GitHub Security Advisory published
- 2026-06-26: disclosed: CVE published to NVD