CVE-2026-95929 in astron-agent
Summary
by MITRE • 09/23/2026
A weakness has been identified in iFlytek astron-agent up to 1.0.7. Affected is an unknown function of the file console/backend/commons/src/main/resources/mapper/ChatBotMarketMapper.xml of the component getBotList API endpoint. Executing a manipulation of the argument sortDirection can lead to sql injection. It is possible to launch the attack remotely. Upgrading to version reward-1575 is able to address this issue. This patch is called 6702be70ae802b1048f5fbec91e690e7b71a4165. You should upgrade the affected component.
If you want to get the best quality for vulnerability data then you always have to consider VulDB.
Analysis
by VulDB Data Team • 09/23/2026
The vulnerability identified in iFlytek astron-agent versions up to 1.0.7 represents a critical security flaw within the ChatBotMarketMapper.xml configuration file, specifically affecting the getBotList API endpoint. This weakness stems from improper neutralization of special elements used in an SQL command, commonly known as SQL injection. The root cause lies in how the application handles user-supplied input for the sortDirection argument without adequate validation or parameterization. When a remote attacker manipulates this specific argument, they can inject malicious SQL code into the backend database queries executed by the system. This type of vulnerability is categorized under CWE-89, which covers Improper Neutralization of Special Elements used in an SQL Command, and it allows for unauthorized access to sensitive data or potential modification of the underlying database structure depending on the privileges granted to the application's database account.
From a technical perspective, the flaw exists because the sortDirection parameter is likely concatenated directly into the SQL query string rather than being passed as a prepared statement parameter. In many ORM frameworks and XML-based mapper configurations like MyBatis, dynamic sorting clauses are often vulnerable if not handled with strict whitelisting or safe parsing logic. By injecting specific payloads through the sortDirection field, an attacker can alter the intended logic of the SQL query. This could lead to Union-based attacks where additional data is retrieved from other tables, error-based extraction that reveals database schema details, or blind boolean-based injections that allow for systematic data exfiltration over time. The ability to execute this attack remotely means that no local access or authentication bypass is necessarily required if the API endpoint is publicly accessible and lacks sufficient input sanitization at the network perimeter level.
The operational impact of this vulnerability is severe, particularly in environments where iFlytek astron-agent is deployed for customer service automation or internal knowledge management systems. Successful exploitation could result in a complete compromise of database integrity and confidentiality. Attackers may extract sensitive user information, proprietary business logic stored within the bot configurations, or administrative credentials if such data resides in accessible tables. Furthermore, depending on the specific SQL dialect and backend permissions, an attacker might execute operating system commands via extended stored procedures or write files to the server disk, leading to full remote code execution. This aligns with ATT&CK technique T1059, Command and Scripting Interpreter, as well as T1078, Valid Accounts, if credential harvesting is part of the attack chain. The persistence of this vulnerability in versions up to 1.0.7 indicates a gap in secure coding practices during the development lifecycle regarding dynamic query construction.
To mitigate this risk, immediate action must be taken by upgrading iFlytek astron-agent to version reward-1575 or later. This specific update includes patch commit 6702be70ae802b1048f5fbec91e690e7b71a4165, which addresses the underlying code flaw by implementing proper input validation and secure query parameterization for the sortDirection argument. In addition to upgrading, organizations should implement defense-in-depth strategies such as deploying a Web Application Firewall configured with rulesets that detect SQL injection patterns in HTTP parameters. Regular security audits and static application security testing focused on XML mapper files can help identify similar vulnerabilities before deployment. It is also advisable to enforce least-privilege principles for the database user account used by the application, ensuring that even if an injection occurs, the attacker cannot perform destructive operations such as dropping tables or modifying system configurations.