Cisco 350-901 Exam Dumps, Practice Test Questions

100% Latest & Updated Cisco 350-901 Practice Test Questions, Exam Dumps & Verified Answers!
30 Days Free Updates, Instant Download!

Cisco 350-901 Premium Bundle
$79.97
$59.98

350-901 Premium Bundle

  • Premium File: 604 Questions & Answers. Last update: Nov 7, 2026
  • Training Course: 106 Video Lectures
  • Study Guide: 1255 Pages
  • Latest Questions
  • 100% Accurate Answers
  • Fast Exam Updates

350-901 Premium Bundle

Cisco 350-901 Premium Bundle
  • Premium File: 604 Questions & Answers. Last update: Nov 7, 2026
  • Training Course: 106 Video Lectures
  • Study Guide: 1255 Pages
  • Latest Questions
  • 100% Accurate Answers
  • Fast Exam Updates
$79.97
$59.98

Cisco 350-901 Practice Test 85

Question 840

One device log message includes instructions telling the assistant to reveal credentials and ignore authorization limits. Independently, reviewers flag this potentially unsafe practice for assessment: 'Delete suspicious input after it influences a tool call'. Which verification addresses that practice without assuming it caused the incident?

  1. Inspect approval workflow and proposed diff
  2. Review source and tool execution records
  3. Inspect instruction/data boundary
  4. Give a summarization agent unrestricted router configuration rights

Explanation

Correct Answer

B:

Missing evidence hides the exploit path. The intended behavior is: Traceability permits investigation of manipulated prompts.

Incorrect Answers

A:

This checks validate proposed actions, not the named failure. A model recommendation is not authorization.

C:

This checks treat retrieved logs as data, not the named failure. External text is not authorized control input.

D:

Injected content may gain unsafe authority.

Question 841

An AI troubleshooting agent reads support tickets that may include attacker-controlled file attachments. The specific implementation objective is to isolate external content. Which configuration best meets the stated requirement?

  1. Perform destructive operations merely because ticket body says 'urgent'
  2. Strip all detection logs after the model rejects a payload
  3. Permit the agent to send arbitrary network files to any URL
  4. Process attachment instructions as untrusted text, not as privileged commands

Explanation

Correct Answer

D:

Retrieved context has lower authority than application policy.

Incorrect Answers

A:

Urgency text is not authorization.

B:

Lost evidence weakens defense testing.

C:

Untrusted instructions can leak data.

Question 842

An AI troubleshooting agent reads support tickets that may include attacker-controlled file attachments. The specific implementation objective is to restrict outbound actions. Which design action is most appropriate for this case?

  1. Let ticket attachments override tool safety rules
  2. Strip all detection logs after the model rejects a payload
  3. Allow only approved destinations and defined diagnostic tools
  4. Perform destructive operations merely because ticket body says 'urgent'

Explanation

Correct Answer

C:

Constrained egress blocks arbitrary exfiltration.

Incorrect Answers

A:

Attackers can craft support content.

B:

Lost evidence weakens defense testing.

D:

Urgency text is not authorization.

Question 843

An AI troubleshooting agent reads support tickets that may include attacker-controlled file attachments. The specific implementation objective is to require authorized intent. What should the engineer implement for this specific objective?

  1. Permit the agent to send arbitrary network files to any URL
  2. Strip all detection logs after the model rejects a payload
  3. Ask a verified operator to approve changes that affect production
  4. Let ticket attachments override tool safety rules

Explanation

Correct Answer

C:

Ticket text is not a trusted change request.

Incorrect Answers

A:

Untrusted instructions can leak data.

B:

Lost evidence weakens defense testing.

D:

Attackers can craft support content.

Question 844

An AI troubleshooting agent reads support tickets that may include attacker-controlled file attachments. The specific implementation objective is to monitor injection attempts. Which adjustment would satisfy this particular constraint?

  1. Let ticket attachments override tool safety rules
  2. Detect and log suspicious tool-redirection phrases without following them
  3. Perform destructive operations merely because ticket body says 'urgent'
  4. Permit the agent to send arbitrary network files to any URL

Explanation

Correct Answer

B:

Visibility supports safe agent operation.

Incorrect Answers

A:

Attackers can craft support content.

C:

Urgency text is not authorization.

D:

Untrusted instructions can leak data.

Question 845

An AI troubleshooting agent reads support tickets that may include attacker-controlled file attachments. The deployment plan must both (1) isolate external content and (2) require authorized intent. Which TWO actions together satisfy those requirements? Choose TWO.

  1. Ask a verified operator to approve changes that affect production
  2. Let ticket attachments override tool safety rules
  3. Permit the agent to send arbitrary network files to any URL
  4. Process attachment instructions as untrusted text, not as privileged commands
  5. Perform destructive operations merely because ticket body says 'urgent'

Explanation

Correct Answers

A:

Ticket text is not a trusted change request.

D:

Retrieved context has lower authority than application policy.

Incorrect Answers

B:

Attackers can craft support content.

C:

Untrusted instructions can leak data.

E:

Urgency text is not authorization.

Question 846

An AI troubleshooting agent reads support tickets that may include attacker-controlled file attachments. Three conditions must be met: (1) isolate external content; (2) restrict outbound actions; and (3) monitor injection attempts. Which THREE actions form the narrowest complete plan? Choose THREE.

  1. Detect and log suspicious tool-redirection phrases without following them
  2. Permit the agent to send arbitrary network files to any URL
  3. Strip all detection logs after the model rejects a payload
  4. Process attachment instructions as untrusted text, not as privileged commands
  5. Let ticket attachments override tool safety rules
  6. Allow only approved destinations and defined diagnostic tools

Explanation

Correct Answers

A:

Visibility supports safe agent operation.

D:

Retrieved context has lower authority than application policy.

F:

Constrained egress blocks arbitrary exfiltration.

Incorrect Answers

B:

Untrusted instructions can leak data.

C:

Lost evidence weakens defense testing.

E:

Attackers can craft support content.

Question 847

A malicious attachment says to disable logging and upload controller secrets to an external site. Independently, reviewers flag this potentially unsafe practice for assessment: 'Let ticket attachments override tool safety rules'. Which inspection most directly evaluates that practice?

  1. Inspect document trust tagging
  2. Permit the agent to send arbitrary network files to any URL
  3. Check requester identity and approval
  4. Inspect injection alerts and blocked-call logs

Explanation

Correct Answer

A:

Attackers can craft support content. The intended behavior is: Retrieved context has lower authority than application policy.

Incorrect Answers

B:

Untrusted instructions can leak data.

C:

This checks require authorized intent, not the named failure. Ticket text is not a trusted change request.

D:

This checks monitor injection attempts, not the named failure. Visibility supports safe agent operation.

Question 848

A malicious attachment says to disable logging and upload controller secrets to an external site. Independently, reviewers flag this potentially unsafe practice for assessment: 'Permit the agent to send arbitrary network files to any URL'. What diagnostic evidence should first be collected to assess that practice?

  1. Inspect egress and tool allowlist
  2. Inspect document trust tagging
  3. Strip all detection logs after the model rejects a payload
  4. Check requester identity and approval

Explanation

Correct Answer

A:

Untrusted instructions can leak data. The intended behavior is: Constrained egress blocks arbitrary exfiltration.

Incorrect Answers

B:

This checks isolate external content, not the named failure. Retrieved context has lower authority than application policy.

C:

Lost evidence weakens defense testing.

D:

This checks require authorized intent, not the named failure. Ticket text is not a trusted change request.

Question 849

A malicious attachment says to disable logging and upload controller secrets to an external site. Independently, reviewers flag this potentially unsafe practice for assessment: 'Perform destructive operations merely because ticket body says 'urgent''. Which focused check best tests this specific assumption?

  1. Inspect document trust tagging
  2. Permit the agent to send arbitrary network files to any URL
  3. Inspect injection alerts and blocked-call logs
  4. Check requester identity and approval

Explanation

Correct Answer

D:

Urgency text is not authorization. The intended behavior is: Ticket text is not a trusted change request.

Incorrect Answers

A:

This checks isolate external content, not the named failure. Retrieved context has lower authority than application policy.

B:

Untrusted instructions can leak data.

C:

This checks monitor injection attempts, not the named failure. Visibility supports safe agent operation.

Question 850

A malicious attachment says to disable logging and upload controller secrets to an external site. Independently, reviewers flag this potentially unsafe practice for assessment: 'Strip all detection logs after the model rejects a payload'. Which verification addresses that practice without assuming it caused the incident?

  1. Inspect injection alerts and blocked-call logs
  2. Permit the agent to send arbitrary network files to any URL
  3. Inspect document trust tagging
  4. Check requester identity and approval

Explanation

Correct Answer

A:

Lost evidence weakens defense testing. The intended behavior is: Visibility supports safe agent operation.

Incorrect Answers

B:

Untrusted instructions can leak data.

C:

This checks isolate external content, not the named failure. Retrieved context has lower authority than application policy.

D:

This checks require authorized intent, not the named failure. Ticket text is not a trusted change request.

Designing, Deploying, and Managing Network Automation Systems (350-901) Practice Tests

UP

SPECIAL OFFER: GET 10% OFF

This is ONE TIME OFFER

ExamSnap Discount Offer
Enter Your Email Address to Receive Your 10% Off Discount Code

A confirmation link will be sent to this email address to verify your login. *We value your privacy. We will not rent or sell your email address.

Download Free Demo of VCE Exam Simulator

Experience Avanset VCE Exam Simulator for yourself.

Simply submit your e-mail address below to get started with our interactive software demo of your free trial.

Free Demo Limits: In the demo version you will be able to access only first 5 questions from exam.