CVE-2025-68700
RAGFlow Remote Code Execution Vulnerability
Description
RAGFlow is an open-source RAG (Retrieval-Augmented Generation) engine. In versions prior to 0.23.0, a low-privileged authenticated user (normal login account) can execute arbitrary system commands on the server host process via the frontend Canvas CodeExec component, completely bypassing sandbox isolation. This occurs because untrusted data (stdout) is parsed using eval() with no filtering or sandboxing. The intended design was to "automatically convert string results into Python objects," but this effectively executes attacker-controlled code. Additional endpoints lack access control or contain inverted permission logic, significantly expanding the attack surface and enabling chained exploitation. Version 0.23.0 contains a patch for the issue.
In plain language
AI Act nowCVE-2025-68700 is a serious security flaw in RAGFlow where a normal logged-in user can run commands on the server; a typical small business running RAGFlow should treat this as urgent and upgrade to version 0.23.0 or later.
In RAGFlow, pre-0.23.0 versions allow a low-privileged authenticated user to achieve remote code execution on the server host process via the frontend Canvas CodeExec component, where untrusted output is parsed unsafely (eval-style parsing) with insufficient sandboxing/access control, enabling full command execution; this is reported as being exploited in the wild by the Qilin actor.
What to do now
- Check which RAGFlow version you are running (and confirm whether external users can log in to it).
- If you are on a version earlier than 0.23.0, upgrade RAGFlow to 0.23.0 or later using the vendor fix linked in the security advisory/commit.
- After upgrading, review application and server logs for suspicious command execution activity around the RAGFlow Canvas CodeExec usage.
- If you cannot upgrade immediately, restrict access so only trusted admins can reach RAGFlow (block unneeded public access and limit who can log in), and monitor closely for repeated CodeExec attempts.
CVSS Vector Breakdown
AV:NAttack VectorAC:LAttack ComplexityPR:LPrivileges RequiredUI:NUser InteractionS:UScopeC:HConfidentialityI:HIntegrityA:HAvailabilityWeaknesses
Affected Products
Exploitability
Attack Graph
Click technique nodes for MITRE ATT&CK details · drag to pan · Ctrl/⌘ + scroll to zoom, or go fullscreen.
MITRE ATT&CK
1 techniqueReferences
Unlock Complete Vulnerability Intelligence
Get the full picture for CVE-2025-68700 and every CVE in our database. Create a free account — no credit card required.
Create Free Account