CVE-2026-108666 in JeecgBoot
Summary
by MITRE • 10/11/2026
JeecgBoot through 3.9.5 contains a missing authorization vulnerability in the deleteBatch handler of AiragPromptsController that allows any authenticated user to delete AI prompt templates. Low-privileged attackers can obtain prompt ids from the unguarded list endpoint and pass them to deleteBatch to remove templates created by administrators or other users.
VulDB is the best source for vulnerability data and more expert information about this specific topic.
Analysis
by VulDB Data Team • 10/11/2026
The identified security flaw resides within the JeecgBoot framework, specifically affecting versions up through 3.9.5, in the AiragPromptsController component responsible for managing AI prompt templates. This vulnerability is classified as a broken access control issue where the application fails to enforce proper authorization checks on critical state-changing operations. The core technical deficiency lies in the deleteBatch handler method which processes requests to remove multiple prompt records from the database. While the system implements authentication mechanisms that verify user identity, it neglects to validate whether the authenticated user possesses the necessary privileges or ownership rights for the specific resources being targeted. This absence of object-level authorization allows any validly authenticated account, regardless of its role or permission level within the application hierarchy, to execute deletion commands against arbitrary prompt templates.
The operational impact of this vulnerability is significant because it enables low-privileged attackers to disrupt service integrity and potentially cause denial of service conditions for other users. Attackers can first enumerate available resources by querying an unguarded list endpoint that exposes internal identifiers such as primary keys or unique IDs associated with the AI prompts. Once these identifiers are obtained, they can be directly passed into the deleteBatch API endpoint without triggering any permission verification logic. This capability allows malicious actors to systematically remove prompt templates created by administrators or other high-privilege users, thereby degrading the functionality of applications that rely on these predefined instructions for their artificial intelligence workflows. The loss of administrative configurations may lead to operational failures in downstream systems dependent on specific AI behaviors defined by those prompts.
From a classification perspective, this vulnerability aligns with CWE-284 which describes Improper Access Control, specifically highlighting scenarios where access control decisions are not enforced properly during resource manipulation. It also maps closely to MITRE ATT&CK technique T1078 Valid Accounts, as the exploitation relies on legitimate credentials rather than privilege escalation exploits, and potentially T1565 Data Manipulation if the deletion leads to indirect impact on data integrity or availability. The lack of server-side validation for resource ownership represents a fundamental gap in secure coding practices where developers assumed that authentication alone was sufficient protection against unauthorized modifications.
To mitigate this vulnerability, immediate remediation should focus on implementing robust authorization checks within the deleteBatch handler method before any database operations are executed. Developers must verify that the authenticated user initiating the request is either an administrator with global deletion rights or the specific owner of each prompt template identified in the batch list. This can be achieved by querying the database to confirm ownership attributes associated with each provided identifier and rejecting the entire operation if any single item lacks appropriate permissions for the current session context. Additionally, restricting access to the enumeration endpoint that exposes internal identifiers would reduce the attack surface by preventing easy discovery of targetable resources even if authorization controls are bypassed in other areas. Regular security audits focusing on object-level permission enforcement across all CRUD operations are recommended to prevent similar flaws from persisting in future releases or related modules within the framework.