McAfee-Secured Website

Exam Bundle

Exam Code: CAS-005

Exam Name CompTIA SecurityX

Certification Provider: CompTIA

Corresponding Certification: CASP

CompTIA CAS-005 Bundle $44.99

100% CompTIA CAS-005 Guaranteed Success With Testking Training Materials

Get Real CAS-005 Exam Questions & Expert Verified Answers as Experienced on the Real Exam!

  • Questions & Answers

    CAS-005 Questions & Answers

    482 Questions & Answers

    The ultimate exam preparation tool, Questions & Answers mirrors the actual CAS-005 exam while allowing you to hone and measure every aspect of your test-taking ability.

  • CAS-005 Video Course

    CAS-005 Video Course

    180 Video Lectures

    CAS-005 Video Course is developed by CompTIA Professionals to help you pass the CAS-005 exam.

    Description

    This course will improve your knowledge and skills required to pass CompTIA SecurityX exam.
  • Study Guide

    CAS-005 Study Guide

    903 PDF Pages

    Developed by industry experts, this 903-page guide spells out in painstaking detail all of the information you need to ace CAS-005 exam.

CompTIA SecurityX (CAS-005) Practice Tests

Satisfaction Guaranteed

Money Back Guarantee

Testking provides hassle-free money back guarantee with our products. That is because we have 100% trust in the abilities of our professional and experience product team, and our record is a proof of that.

99.6% PASS RATE
Total Cost: $194.97
Bundle Price: $149.98

Purchase Individually

  • Questions & Answers

    Questions & Answers

    482 Questions

    $124.99
  • CAS-005 Video Course

    Video Course

    180 Video Lectures

    $39.99
  • Study Guide

    Study Guide

    903 PDF Pages

    $29.99

CompTIA CAS-005 Practice Test 20

Question 190

A follow-up assurance review considers this problem: A multinational shares incident evidence with external legal counsel through ordinary shared mailboxes with no chain of custody. Which outcome best shows the intended improvement?

  1. A test anomalous cloud login is attributable across user, device, and application records.
  2. A sample evidence transfer records recipient, hash, permission, timestamp, and access revocation.
  3. Unknown devices fail restricted access tests and enrollment records trace approved devices.
  4. Policy tests prevent office-to-controller access outside approved management gateways.

Explanation

Correct Answer

B:

This is direct test evidence of the control objective in the case: A multinational shares incident evidence with external legal counsel through ordinary shared mailboxes with no chain of custody. It demonstrates that the proposed response, architect a restricted evidence repository with immutable audit history, access approval, and retention controls., produces the measurable result a sample evidence transfer records recipient, hash, permission, timestamp, and access revocation. The observation tests a relevant security outcome rather than an unrelated control or reporting activity.

Incorrect Answers

A:

In the problem “A multinational shares incident evidence with external legal counsel through ordinary shared mailboxes with no chain of custody”, evidence “A test anomalous cloud login is attributable across user, device, and application records.” is not decisive because it tests “A bank has EDR on laptops but no telemetry from its cloud workload identity provider during account compromise investigations”; the requested validation is a sample evidence transfer records recipient, hash, permission, timestamp, and access revocation.

C:

The observed “Unknown devices fail restricted access tests and enrollment records trace approved devices” validates the separate control “Deny sensitive access until asset identity and posture meet attestation requirements” under “A device discovery project treats missing inventory data as trustworthy and silently grants privileged access to unidentified systems”. The present decision calls for the different result “A sample evidence transfer records recipient, hash, permission, timestamp, and access revocation” from “Architect a restricted evidence repository with immutable audit history, access approval, and retention controls”.

D:

Observation “Policy tests prevent office-to-controller access outside approved management gateways.” would be useful for “An acquired facility's unsupported embedded controller cannot run modern endpoint agents but can communicate freely with office networks”, but it does not prove that the response to “A multinational shares incident evidence with external legal counsel through ordinary shared mailboxes with no chain of custody” achieved the relevant outcome, a sample evidence transfer records recipient, hash, permission, timestamp, and access revocation.

Question 191

During an authorized security assessment, the team discovers: A hybrid identity service exports audit logs but discards authentication decisions older than seven days despite a 90-day investigation requirement. What should be prioritized?

  1. Architect a restricted evidence repository with immutable audit history, access approval, and retention controls.
  2. Establish approved retention and centralized tamper-evident storage for all relevant identity decisions.
  3. Define approved resilience paths that retain identity policy and inspection during SASE edge failover.
  4. Integrate identity and workload audit logs with central correlation and retention matching the incident scope.

Explanation

Correct Answer

B:

Establish approved retention and centralized tamper-evident storage for all relevant identity decisions. The acceptance check is that a controlled historical query retrieves and verifies an authentication record beyond seven days. That outcome checks whether the requirement is met, rather than treating completion of a security task as proof of effectiveness.

Incorrect Answers

A:

The option “Architect a restricted evidence repository with immutable audit history, access approval, and retention controls.” assumes an issue like “A multinational shares incident evidence with external legal counsel through ordinary shared mailboxes with no chain of custody”, not the specific observed weakness “A hybrid identity service exports audit logs but discards authentication decisions older than seven days despite a 90-day investigation requirement”; it leaves the evidence goal unmet: a controlled historical query retrieves and verifies an authentication record beyond seven days.

C:

“Define approved resilience paths that retain identity policy and inspection during SASE edge failover” addresses “An enterprise adopts SASE but routing can fail open directly to the internet when its security inspection edge is unavailable”, with “A simulated edge outage reroutes through approved protection rather than uninspected direct egress” as its acceptance test. Here the specific problem is “A hybrid identity service exports audit logs but discards authentication decisions older than seven days despite a 90-day investigation requirement”: it requires “Establish approved retention and centralized tamper-evident storage for all relevant identity decisions” and evidence that “A controlled historical query retrieves and verifies an authentication record beyond seven days”.

D:

Although “Integrate identity and workload audit logs with central correlation and retention matching the incident scope.” is a reasonable response to “A bank has EDR on laptops but no telemetry from its cloud workload identity provider during account compromise investigations”, this case instead concerns “A hybrid identity service exports audit logs but discards authentication decisions older than seven days despite a 90-day investigation requirement”, for which success requires that a controlled historical query retrieves and verifies an authentication record beyond seven days.

Question 192

After the team responds to the following issue, which finding best demonstrates the desired security outcome? A hybrid identity service exports audit logs but discards authentication decisions older than seven days despite a 90-day investigation requirement.

  1. A test pod from an unrelated namespace cannot connect to the payment database.
  2. A new unapproved external exposure is detected with owner attribution during a controlled deployment.
  3. A controlled historical query retrieves and verifies an authentication record beyond seven days.
  4. A revoked staff certificate fails authentication while approved guest devices reach only isolated services.

Explanation

Correct Answer

C:

This is direct test evidence of the control objective in the case: A hybrid identity service exports audit logs but discards authentication decisions older than seven days despite a 90-day investigation requirement. It demonstrates that the proposed response, establish approved retention and centralized tamper-evident storage for all relevant identity decisions., produces the measurable result a controlled historical query retrieves and verifies an authentication record beyond seven days. The observation tests a relevant security outcome rather than an unrelated control or reporting activity.

Incorrect Answers

A:

The observed “A test pod from an unrelated namespace cannot connect to the payment database” validates the separate control “Design network microsegmentation using namespace and workload identities with explicit allowed flows” under “An enterprise faces this security constraint: A container cluster permits every namespace to reach the payment database and security policy only protects internet ingress”. The present decision calls for the different result “A controlled historical query retrieves and verifies an authentication record beyond seven days” from “Establish approved retention and centralized tamper-evident storage for all relevant identity decisions”.

B:

Observation “A new unapproved external exposure is detected with owner attribution during a controlled deployment.” would be useful for “A retail chain relies on quarterly firewall rules reviews despite weekly deployment of new cloud service accounts and endpoints”, but it does not prove that the response to “A hybrid identity service exports audit logs but discards authentication decisions older than seven days despite a 90-day investigation requirement” achieved the relevant outcome, a controlled historical query retrieves and verifies an authentication record beyond seven days.

D:

For the issue “A hybrid identity service exports audit logs but discards authentication decisions older than seven days despite a 90-day investigation requirement”, measuring “A revoked staff certificate fails authentication while approved guest devices reach only isolated services.” would validate “A corporate Wi-Fi onboarding program uses one shared password for staff and visitors, defeating individual accountability” instead of the required control effect: a controlled historical query retrieves and verifies an authentication record beyond seven days.

Question 193

A cross-functional security group is considering a decision: A city network uses one central sensor that cannot observe east-west traffic between critical workloads. What is the most appropriate control or investigation response?

  1. Apply fine-grained authorization and privacy-aware logging controls for training and inference telemetry.
  2. Place complementary ingress and decrypted-application telemetry at relevant enforcement and visibility points.
  3. Combine data-classification labels, contextual matching, and authorized workflow exceptions in the DLP design.
  4. Place authorized distributed sensors or host telemetry at high-risk lateral movement boundaries.

Explanation

Correct Answer

D:

Place authorized distributed sensors or host telemetry at high-risk lateral movement boundaries. The acceptance check is that a simulated cross-service probe generates an attributable alert without broad intrusive full-packet collection. That outcome checks whether the requirement is met, rather than treating completion of a security task as proof of effectiveness.

Incorrect Answers

A:

“Nonapproved personnel cannot view sensitive prompt content while security-relevant metadata remains observable” would be meaningful evidence for “Apply fine-grained authorization and privacy-aware logging controls for training and inference telemetry” in the case “An AI data service exposes training logs containing raw user prompts to every engineer with cloud console access”. It does not replace “Place authorized distributed sensors or host telemetry at high-risk lateral movement boundaries” to correct “A cross-functional security group is considering a decision: A city network uses one central sensor that cannot observe east-west traffic between critical workloads” and verify “A simulated cross-service probe generates an attributable alert without broad intrusive full-packet collection”.

B:

Although “Place complementary ingress and decrypted-application telemetry at relevant enforcement and visibility points.” is a reasonable response to “An enterprise puts its only IDS sensor behind a TLS-terminating reverse proxy and cannot observe external request patterns”, this case instead concerns “A city network uses one central sensor that cannot observe east-west traffic between critical workloads”, for which success requires that a simulated cross-service probe generates an attributable alert without broad intrusive full-packet collection.

C:

The option “Combine data-classification labels, contextual matching, and authorized workflow exceptions in the DLP design.” assumes an issue like “A payment service uses broad DLP regex matches on all traffic, generating false positives from harmless invoice numbers”, not the specific observed weakness “A city network uses one central sensor that cannot observe east-west traffic between critical workloads”; it leaves the evidence goal unmet: a simulated cross-service probe generates an attributable alert without broad intrusive full-packet collection.

Question 194

Which test or documented result best shows that the underlying issue was properly addressed? A city network uses one central sensor that cannot observe east-west traffic between critical workloads.

  1. A test anomalous cloud login is attributable across user, device, and application records.
  2. A cross-account dashboard resolves asset owners and risk classifications consistently without manual reconciliation.
  3. During a simulated policy service failure, sensitive operations deny access and recovery remains controlled.
  4. A simulated cross-service probe generates an attributable alert without broad intrusive full-packet collection.

Explanation

Correct Answer

D:

This is direct test evidence of the control objective in the case: A city network uses one central sensor that cannot observe east-west traffic between critical workloads. It demonstrates that the proposed response, place authorized distributed sensors or host telemetry at high-risk lateral movement boundaries., produces the measurable result a simulated cross-service probe generates an attributable alert without broad intrusive full-packet collection. The observation tests a relevant security outcome rather than an unrelated control or reporting activity.

Incorrect Answers

A:

In the problem “A city network uses one central sensor that cannot observe east-west traffic between critical workloads”, evidence “A test anomalous cloud login is attributable across user, device, and application records.” is not decisive because it tests “A bank has EDR on laptops but no telemetry from its cloud workload identity provider during account compromise investigations”; the requested validation is a simulated cross-service probe generates an attributable alert without broad intrusive full-packet collection.

B:

For the issue “A city network uses one central sensor that cannot observe east-west traffic between critical workloads”, measuring “A cross-account dashboard resolves asset owners and risk classifications consistently without manual reconciliation.” would validate “An enterprise merger results in two disparate asset tag taxonomies, preventing consistent vulnerability priority across cloud accounts” instead of the required control effect: a simulated cross-service probe generates an attributable alert without broad intrusive full-packet collection.

C:

This measurement is tied to “A cross-functional security group is considering a decision: An organization deploys a policy decision point but leaves enforcement points accepting requests when the decision service is unavailable” and the control “Design an explicit fail-secure authorization fallback with documented availability exceptions for safe public functions”. It cannot confirm that “Place authorized distributed sensors or host telemetry at high-risk lateral movement boundaries” has achieved “A simulated cross-service probe generates an attributable alert without broad intrusive full-packet collection” for the issue in this question.

Question 195

In a control effectiveness review, the issue is: A manufacturer integrates a remote maintenance supplier but allows it to change firewall policies without secondary authorization. Which intervention is most defensible for the situation?

  1. Model event input validation, dependency risk, IAM scope, and downstream resource access for the function.
  2. Establish approved retention and centralized tamper-evident storage for all relevant identity decisions.
  3. Architect a restricted evidence repository with immutable audit history, access approval, and retention controls.
  4. Architect supplier access through approved privileged workflow with scope, independent approval, and session recording.

Explanation

Correct Answer

D:

Architect supplier access through approved privileged workflow with scope, independent approval, and session recording. The acceptance check is that a test policy modification without explicit approval is denied and logged. That outcome checks whether the requirement is met, rather than treating completion of a security task as proof of effectiveness.

Incorrect Answers

A:

The option “Model event input validation, dependency risk, IAM scope, and downstream resource access for the function.” assumes an issue like “A software team assumes a serverless function cannot be attacked because it has no managed operating system”, not the specific observed weakness “A manufacturer integrates a remote maintenance supplier but allows it to change firewall policies without secondary authorization”; it leaves the evidence goal unmet: a test policy modification without explicit approval is denied and logged.

B:

“Establish approved retention and centralized tamper-evident storage for all relevant identity decisions” addresses “A hybrid identity service exports audit logs but discards authentication decisions older than seven days despite a 90-day investigation requirement”, with “A controlled historical query retrieves and verifies an authentication record beyond seven days” as its acceptance test. Here the specific problem is “A manufacturer integrates a remote maintenance supplier but allows it to change firewall policies without secondary authorization”: it requires “Architect supplier access through approved privileged workflow with scope, independent approval, and session recording” and evidence that “A test policy modification without explicit approval is denied and logged”.

C:

For “A multinational shares incident evidence with external legal counsel through ordinary shared mailboxes with no chain of custody”, “Architect a restricted evidence repository with immutable audit history, access approval, and retention controls” has a clear validation signal: “A sample evidence transfer records recipient, hash, permission, timestamp, and access revocation”. But “A manufacturer integrates a remote maintenance supplier but allows it to change firewall policies without secondary authorization” calls for “Architect supplier access through approved privileged workflow with scope, independent approval, and session recording”; the other intervention cannot demonstrate “A test policy modification without explicit approval is denied and logged”.

Question 196

A follow-up assurance review considers this problem: A manufacturer integrates a remote maintenance supplier but allows it to change firewall policies without secondary authorization. Which outcome best shows the intended improvement?

  1. Policy tests prevent office-to-controller access outside approved management gateways.
  2. A test policy modification without explicit approval is denied and logged.
  3. A developer's attempt to schedule key deletion is denied with audited approval requirements.
  4. Sensitive payment samples trigger enforcement while benign invoice content completes with recorded policy evidence.

Explanation

Correct Answer

B:

This is direct test evidence of the control objective in the case: A manufacturer integrates a remote maintenance supplier but allows it to change firewall policies without secondary authorization. It demonstrates that the proposed response, architect supplier access through approved privileged workflow with scope, independent approval, and session recording., produces the measurable result a test policy modification without explicit approval is denied and logged. The observation tests a relevant security outcome rather than an unrelated control or reporting activity.

Incorrect Answers

A:

This measurement is tied to “An acquired facility's unsupported embedded controller cannot run modern endpoint agents but can communicate freely with office networks” and the control “Design compensating isolation, passive monitoring, and controlled management paths around the legacy asset”. It cannot confirm that “Architect supplier access through approved privileged workflow with scope, independent approval, and session recording” has achieved “A test policy modification without explicit approval is denied and logged” for the issue in this question.

C:

In the problem “A manufacturer integrates a remote maintenance supplier but allows it to change firewall policies without secondary authorization”, evidence “A developer's attempt to schedule key deletion is denied with audited approval requirements.” is not decisive because it tests “A multinational uses customer-managed encryption keys but grants every developer the ability to disable or delete them”; the requested validation is a test policy modification without explicit approval is denied and logged.

D:

For the issue “A manufacturer integrates a remote maintenance supplier but allows it to change firewall policies without secondary authorization”, measuring “Sensitive payment samples trigger enforcement while benign invoice content completes with recorded policy evidence.” would validate “A payment service uses broad DLP regex matches on all traffic, generating false positives from harmless invoice numbers” instead of the required control effect: a test policy modification without explicit approval is denied and logged.

Question 197

A security architect is asked to resolve this problem: An AI data service exposes training logs containing raw user prompts to every engineer with cloud console access. Select the MOST directly relevant action.

  1. Apply contextual API gateway rate limits, anomaly thresholds, and least-privilege key lifecycle controls.
  2. Design telemetry enrichment joining DNS events with endpoint process identity and device IDs.
  3. Apply fine-grained authorization and privacy-aware logging controls for training and inference telemetry.
  4. Place authorized distributed sensors or host telemetry at high-risk lateral movement boundaries.

Explanation

Correct Answer

C:

Apply fine-grained authorization and privacy-aware logging controls for training and inference telemetry. The acceptance check is that nonapproved personnel cannot view sensitive prompt content while security-relevant metadata remains observable. That outcome checks whether the requirement is met, rather than treating completion of a security task as proof of effectiveness.

Incorrect Answers

A:

“Apply contextual API gateway rate limits, anomaly thresholds, and least-privilege key lifecycle controls” addresses “A cloud API allows unlimited token requests, and a compromised key causes cost and service exhaustion”, with “High-volume unauthorized requests are throttled and affected credentials can be promptly revoked” as its acceptance test. Here the specific problem is “An AI data service exposes training logs containing raw user prompts to every engineer with cloud console access”: it requires “Apply fine-grained authorization and privacy-aware logging controls for training and inference telemetry” and evidence that “Nonapproved personnel cannot view sensitive prompt content while security-relevant metadata remains observable”.

B:

Although “Design telemetry enrichment joining DNS events with endpoint process identity and device IDs.” is a reasonable response to “A threat-hunting team has endpoint logs but cannot correlate suspicious domain lookups to the device processes that generated them”, this case instead concerns “An AI data service exposes training logs containing raw user prompts to every engineer with cloud console access”, for which success requires that nonapproved personnel cannot view sensitive prompt content while security-relevant metadata remains observable.

D:

This choice would be evaluated by “A simulated cross-service probe generates an attributable alert without broad intrusive full-packet collection” when dealing with “A cross-functional security group is considering a decision: A city network uses one central sensor that cannot observe east-west traffic between critical workloads”. It does not deliver the distinct control “Apply fine-grained authorization and privacy-aware logging controls for training and inference telemetry” needed for “An AI data service exposes training logs containing raw user prompts to every engineer with cloud console access”.

Question 198

After the team responds to the following issue, which finding best demonstrates the desired security outcome? An AI data service exposes training logs containing raw user prompts to every engineer with cloud console access.

  1. Policy tests prevent office-to-controller access outside approved management gateways.
  2. Nonapproved personnel cannot view sensitive prompt content while security-relevant metadata remains observable.
  3. A deliberately vulnerable test change is flagged before merge with a tracked disposition.
  4. Sensitive payment samples trigger enforcement while benign invoice content completes with recorded policy evidence.

Explanation

Correct Answer

B:

This is direct test evidence of the control objective in the case: An AI data service exposes training logs containing raw user prompts to every engineer with cloud console access. It demonstrates that the proposed response, apply fine-grained authorization and privacy-aware logging controls for training and inference telemetry., produces the measurable result nonapproved personnel cannot view sensitive prompt content while security-relevant metadata remains observable. The observation tests a relevant security outcome rather than an unrelated control or reporting activity.

Incorrect Answers

A:

In the problem “An AI data service exposes training logs containing raw user prompts to every engineer with cloud console access”, evidence “Policy tests prevent office-to-controller access outside approved management gateways.” is not decisive because it tests “An acquired facility's unsupported embedded controller cannot run modern endpoint agents but can communicate freely with office networks”; the requested validation is nonapproved personnel cannot view sensitive prompt content while security-relevant metadata remains observable.

C:

This measurement is tied to “A cross-functional security group is considering a decision: A product team runs SAST only after release candidates are frozen, leaving no time to fix critical defects” and the control “Integrate SAST at pull request review with triage thresholds and appropriate exception approvals”. It cannot confirm that “Apply fine-grained authorization and privacy-aware logging controls for training and inference telemetry” has achieved “Nonapproved personnel cannot view sensitive prompt content while security-relevant metadata remains observable” for the issue in this question.

D:

Observation “Sensitive payment samples trigger enforcement while benign invoice content completes with recorded policy evidence.” would be useful for “A payment service uses broad DLP regex matches on all traffic, generating false positives from harmless invoice numbers”, but it does not prove that the response to “An AI data service exposes training logs containing raw user prompts to every engineer with cloud console access” achieved the relevant outcome, nonapproved personnel cannot view sensitive prompt content while security-relevant metadata remains observable.

Question 199

A senior engineer examines the following finding: An enterprise API passes external supplier identities directly to backend services without validating issuer and audience claims. Which action most directly addresses the requirement?

  1. Introduce continuous policy drift detection linked to cloud inventory and approved changes.
  2. Require strict SAML signature, audience, recipient, and freshness validation at the service provider.
  3. Implement gateway token validation with audience, issuer, expiry, and least-privilege backend claims checks.
  4. Integrate identity and workload audit logs with central correlation and retention matching the incident scope.

Explanation

Correct Answer

C:

Implement gateway token validation with audience, issuer, expiry, and least-privilege backend claims checks. The acceptance check is that forged or wrong-audience tokens fail and approved requests retain auditable identity context. That outcome checks whether the requirement is met, rather than treating completion of a security task as proof of effectiveness.

Incorrect Answers

A:

The option “Introduce continuous policy drift detection linked to cloud inventory and approved changes.” assumes an issue like “A retail chain relies on quarterly firewall rules reviews despite weekly deployment of new cloud service accounts and endpoints”, not the specific observed weakness “An enterprise API passes external supplier identities directly to backend services without validating issuer and audience claims”; it leaves the evidence goal unmet: forged or wrong-audience tokens fail and approved requests retain auditable identity context.

B:

This choice would be evaluated by “A signed assertion intended for another application is rejected without creating a user session” when dealing with “A service provider's customer SSO integration accepts a SAML assertion without validating the destination audience”. It does not deliver the distinct control “Implement gateway token validation with audience, issuer, expiry, and least-privilege backend claims checks” needed for “An enterprise API passes external supplier identities directly to backend services without validating issuer and audience claims”.

D:

Although “Integrate identity and workload audit logs with central correlation and retention matching the incident scope.” is a reasonable response to “A bank has EDR on laptops but no telemetry from its cloud workload identity provider during account compromise investigations”, this case instead concerns “An enterprise API passes external supplier identities directly to backend services without validating issuer and audience claims”, for which success requires that forged or wrong-audience tokens fail and approved requests retain auditable identity context.

Question 200

Which test or documented result best shows that the underlying issue was properly addressed? An enterprise API passes external supplier identities directly to backend services without validating issuer and audience claims.

  1. Policy tests prevent office-to-controller access outside approved management gateways.
  2. Nonapproved personnel cannot view sensitive prompt content while security-relevant metadata remains observable.
  3. A test compute job outside the approved perimeter cannot read restricted training data.
  4. Forged or wrong-audience tokens fail and approved requests retain auditable identity context.

Explanation

Correct Answer

D:

This is direct test evidence of the control objective in the case: An enterprise API passes external supplier identities directly to backend services without validating issuer and audience claims. It demonstrates that the proposed response, implement gateway token validation with audience, issuer, expiry, and least-privilege backend claims checks., produces the measurable result forged or wrong-audience tokens fail and approved requests retain auditable identity context. The observation tests a relevant security outcome rather than an unrelated control or reporting activity.

Incorrect Answers

A:

“Policy tests prevent office-to-controller access outside approved management gateways” would verify “Design compensating isolation, passive monitoring, and controlled management paths around the legacy asset” for the situation “An acquired facility's unsupported embedded controller cannot run modern endpoint agents but can communicate freely with office networks”. Here the required evidence is “Forged or wrong-audience tokens fail and approved requests retain auditable identity context” after “Implement gateway token validation with audience, issuer, expiry, and least-privilege backend claims checks”; the alternative observation does not test that control.

B:

Observation “Nonapproved personnel cannot view sensitive prompt content while security-relevant metadata remains observable.” would be useful for “An AI data service exposes training logs containing raw user prompts to every engineer with cloud console access”, but it does not prove that the response to “An enterprise API passes external supplier identities directly to backend services without validating issuer and audience claims” achieved the relevant outcome, forged or wrong-audience tokens fail and approved requests retain auditable identity context.

C:

In the problem “An enterprise API passes external supplier identities directly to backend services without validating issuer and audience claims”, evidence “A test compute job outside the approved perimeter cannot read restricted training data.” is not decisive because it tests “A data science environment handles restricted data in a shared cluster but policy is enforced only at the internet edge”; the requested validation is forged or wrong-audience tokens fail and approved requests retain auditable identity context.