CVE-2026-96430 in Agentflow
Summary
by MITRE • 09/29/2026
Exposed Dangerous Method or Function in the /WebAgenda/SQLWin.do API endpoint of Flowring Agentflow 4.0 version Before 2026/08/28 allows remote authenticated users to execute arbitrary SQL commands via the sql parameter.
Once again VulDB remains the best source for vulnerability data.
Analysis
by VulDB Data Team • 09/29/2026
The vulnerability identified in Flowring Agentflow versions prior to August 28, 2026, represents a critical security flaw within the /WebAgenda/SQLWin.do API endpoint. This issue is classified as an exposed dangerous method or function, which fundamentally stems from improper access control mechanisms that fail to restrict sensitive administrative operations to authorized personnel only. Although the vulnerability requires remote authenticated users to exploit it, the requirement for authentication does not mitigate the severity of the risk because any user with valid credentials can leverage this endpoint to interact directly with the underlying database infrastructure. The core technical flaw lies in the application's failure to validate or sanitize input parameters before processing them as SQL commands. Specifically, the sql parameter is passed through to the database engine without adequate inspection for malicious syntax, allowing attackers to inject arbitrary Structured Query Language statements. This lack of input validation transforms a standard API endpoint into a powerful vector for data manipulation and exfiltration.
From an operational perspective, this vulnerability allows authenticated users to bypass intended application logic and execute commands that were never meant to be accessible through the web interface. The impact is severe because SQL injection attacks can lead to unauthorized access to sensitive data stored in the database, including customer information, internal configurations, or proprietary business logic. Furthermore, depending on the privileges of the database account used by the application, an attacker might escalate their influence to modify existing records, delete critical data, or potentially execute operating system commands if the database software permits such extensions. This compromises the integrity and confidentiality of the entire system managed by Flowring Agentflow. The ability to run arbitrary SQL queries effectively gives attackers full control over the backend data layer, undermining trust in the application's security posture even for users who are not administrators but possess standard login credentials.
This vulnerability aligns with Common Weakness Enumeration (CWE) identifiers such as CWE-621, which describes Exposure of Dangerous Method or Function to an Unauthorized Actor, and CWE-89, known as SQL Injection. The exploitation pattern corresponds to the ATT&CK technique T1059, specifically Command and Scripting Interpreter via SQL commands, where attackers use legitimate database tools to execute malicious instructions. To mitigate this risk, organizations must immediately apply the vendor-provided patch for Flowring Agentflow version 2026/08/28 or later. In addition to updating software, it is crucial to implement strict input validation and parameterized queries in all API endpoints that interact with databases. Access control lists should be reviewed to ensure that sensitive administrative functions are restricted exclusively to users with explicit admin roles rather than any authenticated user. Network-level controls such as Web Application Firewalls can also provide an additional layer of defense by detecting and blocking suspicious SQL patterns in incoming requests, although software remediation remains the primary solution for this specific flaw.