Live › vulnerability › CVE-2026-119337
RAGFlow, an open-source retrieval-augmented generation engine, checks agent access only after the chat completions endpoint has already started running the agent. A logged-in user who knows another user's private agent ID can run that workflow and read its output across account or tenant boundaries.
Summary written from the sources below. Check the originals before acting.
First seen by InfraStack 2026-10-30 20:25 UTC. All times on this page are UTC.
Self-hosted RAGFlow deployments on versions from 0.25.2 up to but not including 0.27.3 are exposed. The risk applies when several users or tenants share the instance and agents are set to private; a single-user deployment has less exposure to the cross-user read.
No product ranges from the sources yet.
0.25.2 through 0.27.30.27.3Each row is the first time that source carried this signal. The analyst’s rows are its own findings.
RAGFlow is an open-source RAG (Retrieval-Augmented Generation) engine. From 0.25.2 until 0.27.3, POST /api/v1/agents/chat/completions invokes completion_openai and returns through the openai-compatible branch before UserCanvasService.accessible() checks whether the caller may access the requested agent. An authenticated user who knows another user's private agent UUID can execute that Only me workflow and read its output across the user or tenant boundary; configured tools may also cause side effects, but stronger impacts were not demonstrated. This issue is fixed in version 0.27.3.
This page as .md · .json · in your agent: get_signal("CVE-2026-119337")
NVD data is provided by NIST and is public domain. Posts quoted belong to their authors.