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 26

Question 250

A follow-up assurance review considers this problem: A Kubernetes service account is mounted automatically into pods that never invoke the cluster API. Which outcome best shows the intended improvement?

  1. A test function accesses approved objects but is denied unrelated project storage.
  2. A test pod without API duties has no mounted token and cannot list unrelated cluster resources.
  3. A crafted event cannot invoke unauthorized actions and vulnerable package versions are detected in the build.
  4. A compromise simulation cannot use third-party service credentials to impersonate unrelated privileged users.

Explanation

Correct Answer

B:

This is direct test evidence of the control objective in the case: A Kubernetes service account is mounted automatically into pods that never invoke the cluster API. It demonstrates that the proposed response, disable unneeded token automount and apply workload-specific RBAC to the remaining service accounts., produces the measurable result a test pod without API duties has no mounted token and cannot list unrelated cluster resources. The observation tests a relevant security outcome rather than an unrelated control or reporting activity.

Incorrect Answers

A:

In the problem “A Kubernetes service account is mounted automatically into pods that never invoke the cluster API”, evidence “A test function accesses approved objects but is denied unrelated project storage.” is not decisive because it tests “A SaaS company allows serverless functions to assume a cloud role with unrestricted storage read permissions across all projects”; the requested validation is a test pod without API duties has no mounted token and cannot list unrelated cluster resources.

C:

The observed “A crafted event cannot invoke unauthorized actions and vulnerable package versions are detected in the build” validates the separate control “Model event input validation, dependency risk, IAM scope, and downstream resource access for the function” under “An enterprise faces this security constraint: A software team assumes a serverless function cannot be attacked because it has no managed operating system”. The present decision calls for the different result “A test pod without API duties has no mounted token and cannot list unrelated cluster resources” from “Disable unneeded token automount and apply workload-specific RBAC to the remaining service accounts”.

D:

Observation “A compromise simulation cannot use third-party service credentials to impersonate unrelated privileged users.” would be useful for “An enterprise uses legacy Kerberos across an acquisition but accepts unconstrained delegation on servers managed by third parties”, but it does not prove that the response to “A Kubernetes service account is mounted automatically into pods that never invoke the cluster API” achieved the relevant outcome, a test pod without API duties has no mounted token and cannot list unrelated cluster resources.

Question 251

In a control effectiveness review, the issue is: A cloud deployment copies service secrets into build logs when a pipeline step fails. Which intervention is most defensible for the situation?

  1. Add location constraints and approved region policies to deployment and data replication workflows.
  2. Treat controller and route changes as policy-sensitive and test enforcement across alternative SD-WAN paths.
  3. Use managed secrets injection and sensitive output redaction for CI/CD workflow execution.
  4. Model event input validation, dependency risk, IAM scope, and downstream resource access for the function.

Explanation

Correct Answer

C:

Use managed secrets injection and sensitive output redaction for CI/CD workflow execution. The acceptance check is that a controlled failing build reveals no usable secret values in the build log. 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 “Add location constraints and approved region policies to deployment and data replication workflows.” assumes an issue like “A cloud team stores application data in a region that meets latency requirements but breaches the signed customer residency agreement”, not the specific observed weakness “A cloud deployment copies service secrets into build logs when a pipeline step fails”; it leaves the evidence goal unmet: a controlled failing build reveals no usable secret values in the build log.

B:

“Treat controller and route changes as policy-sensitive and test enforcement across alternative SD-WAN paths” addresses “A branch SD-WAN rollout changes routing but does not revalidate network security policy after controller failure”, with “A test controller outage maintains permitted paths while prohibited segment communication remains denied” as its acceptance test. Here the specific problem is “A cloud deployment copies service secrets into build logs when a pipeline step fails”: it requires “Use managed secrets injection and sensitive output redaction for CI/CD workflow execution” and evidence that “A controlled failing build reveals no usable secret values in the build log”.

D:

Although “Model event input validation, dependency risk, IAM scope, and downstream resource access for the function.” is a reasonable response to “A software team assumes a serverless function cannot be attacked because it has no managed operating system”, this case instead concerns “A cloud deployment copies service secrets into build logs when a pipeline step fails”, for which success requires that a controlled failing build reveals no usable secret values in the build log.

Question 252

After the team responds to the following issue, which finding best demonstrates the desired security outcome? A cloud deployment copies service secrets into build logs when a pipeline step fails.

  1. A controlled failover starts the standby successfully with a compatible database and acceptable data loss.
  2. A crafted event cannot invoke unauthorized actions and vulnerable package versions are detected in the build.
  3. An unauthorized open ingress rule triggers detection and approved rollback before the next production build.
  4. A controlled failing build reveals no usable secret values in the build log.

Explanation

Correct Answer

D:

This is direct test evidence of the control objective in the case: A cloud deployment copies service secrets into build logs when a pipeline step fails. It demonstrates that the proposed response, use managed secrets injection and sensitive output redaction for CI/CD workflow execution., produces the measurable result a controlled failing build reveals no usable secret values in the build log. The observation tests a relevant security outcome rather than an unrelated control or reporting activity.

Incorrect Answers

A:

The observed “A controlled failover starts the standby successfully with a compatible database and acceptable data loss” validates the separate control “Design configuration and data replication validation before treating standby capacity as recoverable” under “A critical payroll application uses an active-passive cluster whose standby has never had its database schema synchronized”. The present decision calls for the different result “A controlled failing build reveals no usable secret values in the build log” from “Use managed secrets injection and sensitive output redaction for CI/CD workflow execution”.

B:

Observation “A crafted event cannot invoke unauthorized actions and vulnerable package versions are detected in the build.” would be useful for “A software team assumes a serverless function cannot be attacked because it has no managed operating system”, but it does not prove that the response to “A cloud deployment copies service secrets into build logs when a pipeline step fails” achieved the relevant outcome, a controlled failing build reveals no usable secret values in the build log.

C:

In the problem “A cloud deployment copies service secrets into build logs when a pipeline step fails”, evidence “An unauthorized open ingress rule triggers detection and approved rollback before the next production build.” is not decisive because it tests “An infrastructure-as-code pipeline permits manual console changes without alerting owners to network security group drift”; the requested validation is a controlled failing build reveals no usable secret values in the build log.

Question 253

A security architect is asked to resolve this problem: A cloud team stores application data in a region that meets latency requirements but breaches the signed customer residency agreement. Select the MOST directly relevant action.

  1. Enforce trusted artifact signatures and image digest policies before cluster scheduling.
  2. Add drift detection and gated reconciliation against reviewed IaC state for network policies.
  3. Add location constraints and approved region policies to deployment and data replication workflows.
  4. Establish controlled build environments, source integrity checks, and cryptographically signed deployment artifacts.

Explanation

Correct Answer

C:

Add location constraints and approved region policies to deployment and data replication workflows. The acceptance check is that a deployment to a disallowed region fails policy evaluation and permitted replicas stay within contract scope. That outcome checks whether the requirement is met, rather than treating completion of a security task as proof of effectiveness.

Incorrect Answers

A:

For “A container deployment uses a publicly accessible image registry with no signature verification in production admission”, “Enforce trusted artifact signatures and image digest policies before cluster scheduling” has a clear validation signal: “An unsigned or altered image fails admission while the verified release is deployed”. But “A cloud team stores application data in a region that meets latency requirements but breaches the signed customer residency agreement” calls for “Add location constraints and approved region policies to deployment and data replication workflows”; the other intervention cannot demonstrate “A deployment to a disallowed region fails policy evaluation and permitted replicas stay within contract scope”.

B:

Although “Add drift detection and gated reconciliation against reviewed IaC state for network policies.” is a reasonable response to “An infrastructure-as-code pipeline permits manual console changes without alerting owners to network security group drift”, this case instead concerns “A cloud team stores application data in a region that meets latency requirements but breaches the signed customer residency agreement”, for which success requires that a deployment to a disallowed region fails policy evaluation and permitted replicas stay within contract scope.

D:

“Establish controlled build environments, source integrity checks, and cryptographically signed deployment artifacts” addresses “A cloud-native project deploys to production from developers' workstations without reproducible builders”, with “Production accepts only artifacts whose build provenance and signatures validate” as its acceptance test. Here the specific problem is “A cloud team stores application data in a region that meets latency requirements but breaches the signed customer residency agreement”: it requires “Add location constraints and approved region policies to deployment and data replication workflows” and evidence that “A deployment to a disallowed region fails policy evaluation and permitted replicas stay within contract scope”.

Question 254

Which test or documented result best shows that the underlying issue was properly addressed? A cloud team stores application data in a region that meets latency requirements but breaches the signed customer residency agreement.

  1. High-volume unauthorized requests are throttled and affected credentials can be promptly revoked.
  2. A deliberately unsafe redirect change fails before production deployment.
  3. An unsigned or altered image fails admission while the verified release is deployed.
  4. A deployment to a disallowed region fails policy evaluation and permitted replicas stay within contract scope.

Explanation

Correct Answer

D:

This is direct test evidence of the control objective in the case: A cloud team stores application data in a region that meets latency requirements but breaches the signed customer residency agreement. It demonstrates that the proposed response, add location constraints and approved region policies to deployment and data replication workflows., produces the measurable result a deployment to a disallowed region fails policy evaluation and permitted replicas stay within contract scope. The observation tests a relevant security outcome rather than an unrelated control or reporting activity.

Incorrect Answers

A:

In the problem “A cloud team stores application data in a region that meets latency requirements but breaches the signed customer residency agreement”, evidence “High-volume unauthorized requests are throttled and affected credentials can be promptly revoked.” is not decisive because it tests “A cloud API allows unlimited token requests, and a compromised key causes cost and service exhaustion”; the requested validation is a deployment to a disallowed region fails policy evaluation and permitted replicas stay within contract scope.

B:

For the issue “A cloud team stores application data in a region that meets latency requirements but breaches the signed customer residency agreement”, measuring “A deliberately unsafe redirect change fails before production deployment.” would validate “A testing team frequently skips security regression on minor user-interface updates, accidentally reintroducing unsafe redirect handling” instead of the required control effect: a deployment to a disallowed region fails policy evaluation and permitted replicas stay within contract scope.

C:

In “A container deployment uses a publicly accessible image registry with no signature verification in production admission”, “An unsigned or altered image fails admission while the verified release is deployed” is the acceptance signal for “Enforce trusted artifact signatures and image digest policies before cluster scheduling”. The required proof for the selected response “Add location constraints and approved region policies to deployment and data replication workflows” is “A deployment to a disallowed region fails policy evaluation and permitted replicas stay within contract scope”.

Question 255

A senior engineer examines the following finding: A cloud API allows unlimited token requests, and a compromised key causes cost and service exhaustion. Which action most directly addresses the requirement?

  1. Apply contextual API gateway rate limits, anomaly thresholds, and least-privilege key lifecycle controls.
  2. Restrict effective object access, remove public grants, and continuously test external-principal authorization.
  3. Separate key administrators from data users, enforce deletion protections, and require dual approval for destructive key actions.
  4. Implement gateway token validation with audience, issuer, expiry, and least-privilege backend claims checks.

Explanation

Correct Answer

A:

Apply contextual API gateway rate limits, anomaly thresholds, and least-privilege key lifecycle controls. The acceptance check is that high-volume unauthorized requests are throttled and affected credentials can be promptly revoked. That outcome checks whether the requirement is met, rather than treating completion of a security task as proof of effectiveness.

Incorrect Answers

B:

This choice would be evaluated by “An external account cannot list or read protected data and access-control findings clear after verification” when dealing with “A cloud data lake stores confidential exports in a bucket that allows anonymous public listing”. It does not deliver the distinct control “Apply contextual API gateway rate limits, anomaly thresholds, and least-privilege key lifecycle controls” needed for “A cloud API allows unlimited token requests, and a compromised key causes cost and service exhaustion”.

C:

For “An enterprise faces this security constraint: A multinational uses customer-managed encryption keys but grants every developer the ability to disable or delete them”, “Separate key administrators from data users, enforce deletion protections, and require dual approval for destructive key actions” has a clear validation signal: “A developer's attempt to schedule key deletion is denied with audited approval requirements”. But “A cloud API allows unlimited token requests, and a compromised key causes cost and service exhaustion” calls for “Apply contextual API gateway rate limits, anomaly thresholds, and least-privilege key lifecycle controls”; the other intervention cannot demonstrate “High-volume unauthorized requests are throttled and affected credentials can be promptly revoked”.

D:

Although “Implement gateway token validation with audience, issuer, expiry, and least-privilege backend claims checks.” is a reasonable response to “An enterprise API passes external supplier identities directly to backend services without validating issuer and audience claims”, this case instead concerns “A cloud API allows unlimited token requests, and a compromised key causes cost and service exhaustion”, for which success requires that high-volume unauthorized requests are throttled and affected credentials can be promptly revoked.

Question 256

A follow-up assurance review considers this problem: A cloud API allows unlimited token requests, and a compromised key causes cost and service exhaustion. Which outcome best shows the intended improvement?

  1. A crafted event cannot invoke unauthorized actions and vulnerable package versions are detected in the build.
  2. High-volume unauthorized requests are throttled and affected credentials can be promptly revoked.
  3. External callers cannot reach the API; approved private clients authenticate through the intended path.
  4. Malicious test requests are blocked before reaching the application and direct-origin access is denied.

Explanation

Correct Answer

B:

This is direct test evidence of the control objective in the case: A cloud API allows unlimited token requests, and a compromised key causes cost and service exhaustion. It demonstrates that the proposed response, apply contextual API gateway rate limits, anomaly thresholds, and least-privilege key lifecycle controls., produces the measurable result high-volume unauthorized requests are throttled and affected credentials can be promptly revoked. The observation tests a relevant security outcome rather than an unrelated control or reporting activity.

Incorrect Answers

A:

The observed “A crafted event cannot invoke unauthorized actions and vulnerable package versions are detected in the build” validates the separate control “Model event input validation, dependency risk, IAM scope, and downstream resource access for the function” under “An enterprise faces this security constraint: A software team assumes a serverless function cannot be attacked because it has no managed operating system”. The present decision calls for the different result “High-volume unauthorized requests are throttled and affected credentials can be promptly revoked” from “Apply contextual API gateway rate limits, anomaly thresholds, and least-privilege key lifecycle controls”.

C:

In the problem “A cloud API allows unlimited token requests, and a compromised key causes cost and service exhaustion”, evidence “External callers cannot reach the API; approved private clients authenticate through the intended path.” is not decisive because it tests “A cloud integration exposes a private API through an unrestricted public endpoint rather than the expected private service connection”; the requested validation is high-volume unauthorized requests are throttled and affected credentials can be promptly revoked.

D:

For the issue “A cloud API allows unlimited token requests, and a compromised key causes cost and service exhaustion”, measuring “Malicious test requests are blocked before reaching the application and direct-origin access is denied.” would validate “A public utility adds a WAF directly behind the application server, so it cannot reject malformed requests before application processing” instead of the required control effect: high-volume unauthorized requests are throttled and affected credentials can be promptly revoked.

Question 257

An enterprise faces this security constraint: A software team assumes a serverless function cannot be attacked because it has no managed operating system. Which proposal best addresses the stated risk?

  1. Restrict each function's role to its required resources and actions using reviewed cloud IAM policies.
  2. Architect supplier access through approved privileged workflow with scope, independent approval, and session recording.
  3. Model event input validation, dependency risk, IAM scope, and downstream resource access for the function.
  4. Disable unneeded token automount and apply workload-specific RBAC to the remaining service accounts.

Explanation

Correct Answer

C:

Model event input validation, dependency risk, IAM scope, and downstream resource access for the function. The acceptance check is that a crafted event cannot invoke unauthorized actions and vulnerable package versions are detected in the build. That outcome checks whether the requirement is met, rather than treating completion of a security task as proof of effectiveness.

Incorrect Answers

A:

The proposal “Restrict each function's role to its required resources and actions using reviewed cloud IAM policies” manages “A SaaS company allows serverless functions to assume a cloud role with unrestricted storage read permissions across all projects”, where successful operation means “A test function accesses approved objects but is denied unrelated project storage”. The present issue instead needs “Model event input validation, dependency risk, IAM scope, and downstream resource access for the function” to establish “A crafted event cannot invoke unauthorized actions and vulnerable package versions are detected in the build”.

B:

Although “Architect supplier access through approved privileged workflow with scope, independent approval, and session recording.” is a reasonable response to “A manufacturer integrates a remote maintenance supplier but allows it to change firewall policies without secondary authorization”, this case instead concerns “A software team assumes a serverless function cannot be attacked because it has no managed operating system”, for which success requires that a crafted event cannot invoke unauthorized actions and vulnerable package versions are detected in the build.

D:

This choice would be evaluated by “A test pod without API duties has no mounted token and cannot list unrelated cluster resources” when dealing with “A cross-functional security group is considering a decision: A Kubernetes service account is mounted automatically into pods that never invoke the cluster API”. It does not deliver the distinct control “Model event input validation, dependency risk, IAM scope, and downstream resource access for the function” needed for “An enterprise faces this security constraint: A software team assumes a serverless function cannot be attacked because it has no managed operating system”.

Question 258

After the team responds to the following issue, which finding best demonstrates the desired security outcome? A software team assumes a serverless function cannot be attacked because it has no managed operating system.

  1. An unauthorized open ingress rule triggers detection and approved rollback before the next production build.
  2. A controlled failing build reveals no usable secret values in the build log.
  3. A known vulnerable base-layer package triggers a pre-deployment remediation gate.
  4. A crafted event cannot invoke unauthorized actions and vulnerable package versions are detected in the build.

Explanation

Correct Answer

D:

This is direct test evidence of the control objective in the case: A software team assumes a serverless function cannot be attacked because it has no managed operating system. It demonstrates that the proposed response, model event input validation, dependency risk, IAM scope, and downstream resource access for the function., produces the measurable result a crafted event cannot invoke unauthorized actions and vulnerable package versions are detected in the build. The observation tests a relevant security outcome rather than an unrelated control or reporting activity.

Incorrect Answers

A:

In the problem “A software team assumes a serverless function cannot be attacked because it has no managed operating system”, evidence “An unauthorized open ingress rule triggers detection and approved rollback before the next production build.” is not decisive because it tests “An infrastructure-as-code pipeline permits manual console changes without alerting owners to network security group drift”; the requested validation is a crafted event cannot invoke unauthorized actions and vulnerable package versions are detected in the build.

B:

For the issue “A software team assumes a serverless function cannot be attacked because it has no managed operating system”, measuring “A controlled failing build reveals no usable secret values in the build log.” would validate “A cloud deployment copies service secrets into build logs when a pipeline step fails” instead of the required control effect: a crafted event cannot invoke unauthorized actions and vulnerable package versions are detected in the build.

C:

In “An application uses dynamic code scanning but never tests vulnerable open-source libraries in container base images”, “A known vulnerable base-layer package triggers a pre-deployment remediation gate” is the acceptance signal for “Add software composition and container image vulnerability evaluation with dependency ownership”. The required proof for the selected response “Model event input validation, dependency risk, IAM scope, and downstream resource access for the function” is “A crafted event cannot invoke unauthorized actions and vulnerable package versions are detected in the build”.

Question 259

An enterprise security review reports: A zero-trust network project authenticates employees at the VPN gateway but then grants unrestricted access to every internal application. Which response best addresses the underlying security issue?

  1. Introduce attribute-based authorization using jurisdiction, data classification, role, and contextual conditions.
  2. Model subject-to-object authorization at resource level with audience-bound token claims.
  3. Adopt time-bound subject-to-resource grants and continuous authorization checks for contractor sessions.
  4. Enforce per-application authorization based on verified subject, device, and resource context after network connection.

Explanation

Correct Answer

D:

Enforce per-application authorization based on verified subject, device, and resource context after network connection. The acceptance check is that an authorized user can reach their application but cannot access unrelated protected workloads through VPN. That outcome checks whether the requirement is met, rather than treating completion of a security task as proof of effectiveness.

Incorrect Answers

A:

This choice would be evaluated by “Cross-region attempts to retrieve restricted records fail while permitted local access succeeds” when dealing with “An application grants the same permissions to users in different regions despite local data-residency restrictions”. It does not deliver the distinct control “Enforce per-application authorization based on verified subject, device, and resource context after network connection” needed for “A zero-trust network project authenticates employees at the VPN gateway but then grants unrestricted access to every internal application”.

B:

For “A cross-functional security group is considering a decision: An API product's zero-trust plan does not define the relationship between a third-party integration token and individual customer records”, “Model subject-to-object authorization at resource level with audience-bound token claims” has a clear validation signal: “A partner token for one customer cannot retrieve another customer's records”. But “A zero-trust network project authenticates employees at the VPN gateway but then grants unrestricted access to every internal application” calls for “Enforce per-application authorization based on verified subject, device, and resource context after network connection”; the other intervention cannot demonstrate “An authorized user can reach their application but cannot access unrelated protected workloads through VPN”.

C:

For “A company grants temporary external contractors standing application access because the VPN login remains valid for 30 days”, “Adopt time-bound subject-to-resource grants and continuous authorization checks for contractor sessions” has a clear validation signal: “Contract expiry ends access to the protected application even before the VPN session expires”. But “A zero-trust network project authenticates employees at the VPN gateway but then grants unrestricted access to every internal application” calls for “Enforce per-application authorization based on verified subject, device, and resource context after network connection”; the other intervention cannot demonstrate “An authorized user can reach their application but cannot access unrelated protected workloads through VPN”.

Question 260

Which test or documented result best shows that the underlying issue was properly addressed? A zero-trust network project authenticates employees at the VPN gateway but then grants unrestricted access to every internal application.

  1. A simulated high-risk session triggers additional verification or denial without blocking normal authorized sessions.
  2. External callers cannot reach the API; approved private clients authenticate through the intended path.
  3. An authorized user can reach their application but cannot access unrelated protected workloads through VPN.
  4. A test compute job outside the approved perimeter cannot read restricted training data.

Explanation

Correct Answer

C:

This is direct test evidence of the control objective in the case: A zero-trust network project authenticates employees at the VPN gateway but then grants unrestricted access to every internal application. It demonstrates that the proposed response, enforce per-application authorization based on verified subject, device, and resource context after network connection., produces the measurable result an authorized user can reach their application but cannot access unrelated protected workloads through VPN. The observation tests a relevant security outcome rather than an unrelated control or reporting activity.

Incorrect Answers

A:

In “A remote access architecture relies on a fixed hardware token and ignores risky sign-ins from novel geography and unmanaged devices”, “A simulated high-risk session triggers additional verification or denial without blocking normal authorized sessions” is the acceptance signal for “Add conditional access evaluating device posture, location anomalies, and transaction sensitivity”. The required proof for the selected response “Enforce per-application authorization based on verified subject, device, and resource context after network connection” is “An authorized user can reach their application but cannot access unrelated protected workloads through VPN”.

B:

Observation “External callers cannot reach the API; approved private clients authenticate through the intended path.” would be useful for “A cloud integration exposes a private API through an unrestricted public endpoint rather than the expected private service connection”, but it does not prove that the response to “A zero-trust network project authenticates employees at the VPN gateway but then grants unrestricted access to every internal application” achieved the relevant outcome, an authorized user can reach their application but cannot access unrelated protected workloads through VPN.

D:

For the issue “A zero-trust network project authenticates employees at the VPN gateway but then grants unrestricted access to every internal application”, measuring “A test compute job outside the approved perimeter cannot read restricted training data.” would validate “A data science environment handles restricted data in a shared cluster but policy is enforced only at the internet edge” instead of the required control effect: an authorized user can reach their application but cannot access unrelated protected workloads through VPN.