RAGFlow: OpenAI-compatible agent completion bypasses private canvas authorization
Affects RAGFlow 0.25.2 through 0.27.2 when agents are called via the OpenAI-compatible completions endpoint. The vendor gives no separate rating, and the fix is in 0.27.3.
- Published
- Last modified
- Where the record stands: published, rejected by its publisher, or superseded by another id.What do these mean? →
- published
- Whether CISA lists it in its Known Exploited Vulnerabilities catalog: proof that attackers are using it.What do these mean? →
- Not listed in CISA KEV
- How many real-world setups can hit the bug, from the vendor's advisory and CISA data. It does not know your setup.What do these mean? →
- Specific setups only
- Fix
- Fixed in 0.27.3
- Whether working attack code is public (Exploit-DB, Metasploit, Nuclei or GitHub). Public code makes attacks easier.What do these mean? →
- None on Exploit-DB, Metasploit or GitHub
Summary
RAGFlow, an open-source retrieval-augmented generation engine, is affected through its OpenAI-compatible agent chat completions API. An authenticated user who knows another user's private agent ID can run that agent and read its output, bypassing the access controls meant to keep it private.
Summarized by anthropic/claude-haiku-5.5. Verify details against the original sources.
Read the original description
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.
Timeline
- CVE published
- First seen by cve.now
Affected versions
- infiniflow ragflow
From 0.25.2 · fixed in 0.27.3
References 4
GitHub · infiniflow/ragflow
Questions about CVE-2026-119337
What does this mean?
A private RAGFlow agent, one with permission set to Only me, can be run through the OpenAI-compatible agent completion endpoint without the authorization check that should block access to it. The caller must know the agent's UUID.Source: github.com
Am I affected, and how do I check?
RAGFlow versions 0.25.2 through 0.27.2 are listed as affected. The issue applies to requests to POST /api/v1/agents/chat/completions that set "openai-compatible": true and target an agent with permission Only me. To check, look at the RAGFlow version the deployment runs and whether this endpoint is used with the openai-compatible flag.Source: github.com
What should I do about it?
Upgrade RAGFlow to version 0.27.3, which is the listed fixed version.Source: github.com
Is it being exploited?
It is not in CISA's Known Exploited Vulnerabilities catalog, and CISA's SSVC assessment does not mark exploitation as active. That does not mean it has not been attacked: vendors and researchers can report attacks before CISA lists them. No public exploit code is tracked.Source: cve.now feed data