Description

Unauthenticated Cross Site Scripting (XSS) in AI Engine <= 3.6.8 versions.

Executive Summary

Technical Details

Mitigations

References

Additional Metadata

{
  "score": "7.1",
  "pubDate": "2026-08-06T15:17:16.360Z",
  "pubdate": "2026-08-06T15:17:16.360Z",
  "executiveSummary": "An unauthenticated Cross-Site Scripting (XSS) vulnerability has been identified in the AI Engine product. This security flaw enables remote, unauthenticated attackers to execute arbitrary client-side scripts, such as JavaScript, within the context of a victim's browser session. The vulnerability affects all versions of AI Engine up to and including version 3.6.8.\nThe primary risk implication of this vulnerability is the potential compromise of user sessions, potential theft of sensitive session tokens or cookies, and unauthorized execution of actions within the web application on behalf of the victim. Because the vulnerability is unauthenticated, no prior access or credentials are required by the attacker to initiate the exploit sequence.\nSuccessful exploitation requires the victim to interact with or load malicious inputs crafted by the attacker, leading to the rendering of unsanitized data in the application's interface. Organizations utilizing vulnerable versions of AI Engine face risks of defacement, session hijacking, and redirection to malicious external sites.",
  "technicalDetails": "The vulnerability resides in the input handling and output rendering mechanisms of the AI Engine product in versions 3.6.8 and prior. Specifically, the application fails to properly sanitize or encode user-supplied input before reflecting or rendering it back to the client browser. This lack of context-aware output encoding constitutes the fundamental root cause of the Cross-Site Scripting (XSS) condition.\nAs an unauthenticated vulnerability, network exposure is inherent to any publicly accessible deployment of the affected product. Attackers do not require any privilege requirements or authentication credentials to deliver the malicious payload. The attack vector typically involves crafting a malicious URL or injecting malicious payloads into input fields that are subsequently processed and rendered by the application.\nThe attack flow proceeds as follows: First, the unauthenticated attacker crafts a malicious payload containing executable JavaScript encapsulated within HTML tags. Second, the attacker induces a victim to interact with the malicious input, either through direct interaction or via reflected parameters handled by the vulnerable component. Third, the application processes the request and incorporates the unsanitized payload into the HTTP response returned to the client. Finally, the victim's browser parses the response, executing the embedded script within the security context of the victim's session.\nThe payload behavior includes executing arbitrary script logic in the victim's browser, which can lead to post-exploitation impacts such as DOM manipulation, session token exfiltration, keylogging, or performing unauthorized administrative actions if an authenticated user views the malicious content. The vulnerable component lacks proper input validation filters and strict Content Security Policy (CSP) enforcement to neutralize injected scripts."
}