Using AI for Web Exploitation GOAA Practice Questions (Page 2)

Questions 6–10

1. 6expert · hard

   A red team is using an AI-driven orchestration tool to automate the exploitation of a web application. The tool has identified a stored XSS vulnerability in a comment field and a CSRF vulnerability in the password change form. The team wants to chain these vulnerabilities to take over an administrator account. The administrator is known to visit the comment page regularly. Which exploitation chain is most effective?

   • AInject a stored XSS payload that automatically submits a CSRF request to change the administrator's password when the admin views the comment.
   • BUse the stored XSS to deface the comment page and hope the administrator resets the password.
   • CUse the CSRF vulnerability to change the administrator's password, then use the stored XSS to steal the session cookie.
   • DUse the stored XSS to steal the administrator's session cookie, then use the CSRF vulnerability to change the password.
   Reveal Solution Discussion Select an answer first
2. 7expert · medium

   A red team is building an automated exploitation workflow for a web application that has both a staging and a production environment. The workflow must run against staging first, and only proceed to production if the staging run is successful and the client approves. The workflow must also handle the case where the AI tool generates a payload that works in staging but fails in production due to environment differences. What is the best design?

   • AUse the same AI-generated payload for both environments without re-validation, because AI is reliable.
   • BImplement a workflow with a human-in-the-loop approval gate between staging and production, and include a rollback mechanism that reverts any changes made in production if the payload fails.
   • CRun the workflow against production directly to save time, since the staging environment is identical.
   • DAutomate the approval process by having the AI tool self-approve based on its confidence score.
   Reveal Solution Discussion Select an answer first
4. 8application · medium

   A red team is automating the exploitation of a web application using an AI-driven orchestration tool. The tool is configured to chain multiple vulnerabilities: first, an SQL injection to extract user credentials, then a session fixation attack to impersonate a user. The team wants to ensure that the entire chain is executed reliably. Which configuration is most important?

   • ADisable error handling so that the tool continues even if a step fails.
   • BDefine dependencies between the steps so that the session fixation attack only runs after the SQL injection successfully extracts credentials.
   • CRun all steps in parallel to save time and reduce the overall attack duration.
   • DUse the same payload for both steps to simplify the configuration.
   Reveal Solution Discussion Select an answer first
5. 9foundation · medium

   What is a common input source that AI-based vulnerability discovery tools use to identify potential web application flaws?

   • AThe brand of the web application firewall
   • BThe organization's employee email addresses
   • CHTTP request and response traffic logs
   • DThe physical location of the web server
   Reveal Solution Discussion Select an answer first
6. 10expert · medium

   A penetration testing firm is using an AI-assisted vulnerability scanner to assess a client's web application. The scanner uses a machine learning model to identify potential vulnerabilities. The client requires that the scanner not send any exploit payloads to the production environment, only benign requests. The scanner has identified a potential SQL injection point. What is the best way to validate the finding without violating the constraint?

   • ASend a benign request to the production endpoint and observe if the application behaves differently, without any payload.
   • BUse the AI model to generate a non-destructive boolean-based payload that only tests for a true/false response, and run it against a staging copy of the application.
   • CAsk the AI model to provide a confidence score and treat scores above 90% as confirmed vulnerabilities.
   • DRun a full SQL injection exploit against production and rely on the WAF to block it if it is not a real vulnerability.
   Reveal Solution Discussion Select an answer first