Question 620
An external threat feed description includes instructions telling a summarization assistant to delete the security case. Which investigative step would most directly establish the relevant scope?
- Identify sensitive attributes and where the prompt was retained
- Test the playbook with benign edge cases and critical identities
- Replay known malicious fixtures against old and new models
- Separate cited threat facts from operational commands in the feed
Explanation
Correct Answer
D:Given “An external threat feed description includes instructions telling a summarization assistant to delete the security case”, this distinction prevents untrusted descriptions from controlling case disposition; that best addresses scope analysis in Prompt injection in threat enrichment.
Incorrect Answers
A:The choice is relevant to Sensitive data entered into AI copilot, where response scope depends on what data left approved systems and who may access it; for scope analysis in Prompt injection in threat enrichment, this distinction prevents untrusted descriptions from controlling case disposition.
B:For scope analysis in Prompt injection in threat enrichment, Automated response approval boundary is a different mechanism: simulation demonstrates whether control logic causes disproportionate disruption; here, this distinction prevents untrusted descriptions from controlling case disposition.
C:The choice is relevant to AI alert score drift, where comparative testing distinguishes benign seasonal variation from dangerous detection drift; for scope analysis in Prompt injection in threat enrichment, this distinction prevents untrusted descriptions from controlling case disposition.
Question 621
A low-confidence endpoint detection fires on a server containing sensitive identity keys and remains unassigned. What security condition most directly explains the pattern?
- Asset-criticality-adjusted alert triage
- Cross-sensor severity reconciliation
- Duplicate incident generation from overlapping rules
- Authorized-change-driven false positive
Explanation
Correct Answer
A:Given “A low-confidence endpoint detection fires on a server containing sensitive identity keys and remains unassigned”, low model confidence can still warrant fast examination on a high-impact asset; that best addresses classification in Low-confidence alert with crown-jewel impact.
Incorrect Answers
B:The choice is relevant to Conflicting alert severity from two sources, where device ratings differ because sensors observe different stages of possible exploitation; for classification in Low-confidence alert with crown-jewel impact, low model confidence can still warrant fast examination on a high-impact asset.
C:For classification in Low-confidence alert with crown-jewel impact, Repeated duplicate detection tickets is a different mechanism: independent detectors may observe one event chain but should not create redundant investigations; here, low model confidence can still warrant fast examination on a high-impact asset.
D:The choice is relevant to Benign admin activity resembling attack, where the alert may correctly detect behavior but misclassify its approved operational context; for classification in Low-confidence alert with crown-jewel impact, low model confidence can still warrant fast examination on a high-impact asset.
Question 622
A new user Run-key value points to an executable in a roaming profile and starts after every interactive sign-in. Which engineering change provides the most relevant ongoing visibility?
- Correlate unusual LSASS handle access with new credential use
- Join suspicious PowerShell arguments with rare destinations
- Alert on autoruns pointing to user-writable paths
- Alert on signed loaders executing user-writable modules
- Alert on security-setting changes followed by heartbeat gaps
Explanation
Correct Answer
C:The Persistence through registry Run key evidence “A new user Run-key value points to an executable in a roaming profile and starts after every interactive sign-in” is decisive for ongoing detection improvement because signer and prevalence context reduce false positives from legitimate updaters.
Incorrect Answers
A:The choice is relevant to Credential store memory access, where sequence-based rules distinguish routine observation from suspected theft; for ongoing detection improvement in Persistence through registry Run key, signer and prevalence context reduce false positives from legitimate updaters.
B:For ongoing detection improvement in Persistence through registry Run key, Encoded PowerShell execution chain is a different mechanism: correlated execution and egress cuts noise from benign automation; here, signer and prevalence context reduce false positives from legitimate updaters.
D:For ongoing detection improvement in Persistence through registry Run key, Rundll32 unusual DLL loading is a different mechanism: path and signer combinations improve precision over executable-name matching; here, signer and prevalence context reduce false positives from legitimate updaters.
E:The choice is relevant to Defender telemetry suppression attempt, where state transitions reveal suppression that signature detections alone may miss; for ongoing detection improvement in Persistence through registry Run key, signer and prevalence context reduce false positives from legitimate updaters.
Question 623
A newly created customer account retrieves tens of thousands of profiles through a paginated API overnight. The finding has been corroborated and the response team has authorization to act. Which targeted action should the authorized response team prioritize?
- Revoke the exposed credential and replace it using a secrets manager
- Suspend unauthorized automated access and preserve transaction records
- Disable the vulnerable function and deploy parameterized database access
- Remove the web shell after evidence capture and rebuild if integrity is uncertain
Explanation
Correct Answer
B:Containing the caller limits further extraction while evidence is retained; this is the relevant response decision mechanism for Automated scraping of protected records given “A newly created customer account retrieves tens of thousands of profiles through a paginated API overnight”.
Incorrect Answers
A:For the alternative Leaked application database password situation, secret invalidation prevents reuse while application dependencies are migrated; by contrast, for Automated scraping of protected records response decision, containing the caller limits further extraction while evidence is retained.
C:For the alternative Confirmed SQL injection data exposure situation, temporary exposure reduction and code repair are both necessary; by contrast, for Automated scraping of protected records response decision, containing the caller limits further extraction while evidence is retained.
D:In Web-shell persistence after patching, response must eradicate the foothold and verify system trust; the reported Automated scraping of protected records case calls for a different response decision conclusion because containing the caller limits further extraction while evidence is retained.
Question 624
A detection engineer excludes an entire server subnet to reduce a noisy administrative tool alert. Which scoping query or comparison best tests the working hypothesis?
- Replay positive and negative tests against the exclusion
- Rebuild distribution by peer cohort and maintenance period
- Run known-positive and known-negative events through the new rule
- Enforce order and maximum time gaps during rule replay
- Recompute counts with overlapping or rolling windows
Explanation
Correct Answer
A:Given “A detection engineer excludes an entire server subnet to reduce a noisy administrative tool alert”, known attack simulations verify whether the exception masks true detections; that best addresses scope analysis in Overbroad allowlist suppression.
Incorrect Answers
B:That option fits Peer-group baseline mismatch since comparable entities provide a more defensible anomaly threshold; it does not fit the Overbroad allowlist suppression scope analysis scenario, where known attack simulations verify whether the exception masks true detections.
C:This option addresses Detection-as-code regression because test fixtures demonstrate detection behavior after each change; the Overbroad allowlist suppression scope analysis question instead turns on the fact that known attack simulations verify whether the exception masks true detections.
D:That option fits Multi-stage attack sequence correlation since a valid sequence should not match unrelated events spread across weeks; it does not fit the Overbroad allowlist suppression scope analysis scenario, where known attack simulations verify whether the exception masks true detections.
E:This option addresses Sliding-window aggregation error because the test shows whether the intended threshold is met despite bucket splits; the Overbroad allowlist suppression scope analysis question instead turns on the fact that known attack simulations verify whether the exception masks true detections.
Question 625
A helpdesk user exports many infrastructure passwords from a privileged-access vault without a recorded emergency. The finding has been corroborated and the response team has authorization to act. Which operational remedy is best aligned to the verified issue?
- Restrict unapproved access under coordinated legal and HR procedures
- Route intrusive collection requests through legal and HR authorization
- Freeze the compromised workflow and restore validated records
- Revoke suspicious vault access and rotate potentially exposed secrets
Explanation
Correct Answer
D:The Privileged vault secret retrieval anomaly evidence “A helpdesk user exports many infrastructure passwords from a privileged-access vault without a recorded emergency” is decisive for response decision because response must address both vault entitlement and secrets already retrieved.
Incorrect Answers
A:This option addresses Suspected employee intellectual-property export because preserve evidence and ensure proportionate action before final conclusions; the Privileged vault secret retrieval anomaly response decision question instead turns on the fact that response must address both vault entitlement and secrets already retrieved.
B:That option fits Employee privacy during insider case since controlled approvals protect the investigation and employee rights; it does not fit the Privileged vault secret retrieval anomaly response decision scenario, where response must address both vault entitlement and secrets already retrieved.
C:This option addresses Unauthorized change to payroll records because controlled correction avoids further corruption while preserving original evidence; the Privileged vault secret retrieval anomaly response decision question instead turns on the fact that response must address both vault entitlement and secrets already retrieved.
Question 626
One previously unseen POST endpoint triggers web-server child processes that invoke a command shell. What monitoring improvement would best help detect recurrence?
- Detect allowed high-risk payloads with backend effect
- Alert on web-worker processes spawning shells
- Alert on repeated token failures toward privileged resources
- Alert on privileged grants outside authorized workflows
Explanation
Correct Answer
B:In the Web-shell request and process correlation case, parent-child linkage has more value than scanning page names alone; the observation “One previously unseen POST endpoint triggers web-server child processes that invoke a command shell” makes this ongoing detection improvement choice the fitting one.
Incorrect Answers
A:This option addresses WAF bypass with origin success because joint edge-origin rules are stronger than WAF signature counts alone; the Web-shell request and process correlation ongoing detection improvement question instead turns on the fact that parent-child linkage has more value than scanning page names alone.
C:This option addresses API bearer-token guessing because sensitive path context raises priority above generic 401 volume; the Web-shell request and process correlation ongoing detection improvement question instead turns on the fact that parent-child linkage has more value than scanning page names alone.
D:That option fits Unexpected web administrative role assignment since role-sensitive detections should include actor entitlement and change context; it does not fit the Web-shell request and process correlation ongoing detection improvement scenario, where parent-child linkage has more value than scanning page names alone.
Question 627
One public address attempts the same password against hundreds of employees. Which THREE considerations should appear in an analyst's confidence assessment? Choose THREE.
- One originating IP can represent a proxy or shared network, not necessarily one physical device.
- A spray pattern proves all attempted accounts have been compromised.
- A lack of successful sign-ins in the queried source does not exclude successes in unqueried systems.
- The same password across users automatically means a benign enterprise password test.
- Successes that occur later from different IPs may still be related and need correlation.
- Enforcing multifactor authentication makes an attempted password spray impossible.
Explanation
Correct Answers
A:NAT, cloud gateways and shared infrastructure may aggregate different clients. Source attribution requires provider, device and session context rather than treating one IP as one endpoint.
C:Authentication logs may be incomplete across identity providers, legacy protocols or the selected period. The report should describe which sources and time windows were checked.
E:Distributed access and follow-on authentication can conceal a direct source-to-success link. User, device, token and timing pivots help test whether targeted accounts were later used.
Incorrect Answers
B:Failed authentication attempts are evidence of targeting, not proof of access. Successful sign-ins, token issuance or other account activity are needed to identify actual compromise.
D:Pattern alone does not demonstrate authorization. Change records, test source identity and scope must be confirmed before treating a password spray as sanctioned activity.
F:MFA can reduce the chance that guessed passwords lead to access, but it does not stop attackers attempting passwords or exploiting exceptions and legacy flows.
Question 628
A serverless workload begins contacting an unapproved Internet host after a dependency update. The finding has been corroborated and the response team has authorization to act. Which targeted action should the authorized response team prioritize?
- Assign remediation owners for validated exposure findings
- Revoke compromised federation sessions and affected provider grants
- Revoke confirmed abusive access and correct IAM bindings
- Deploy tested detection rules and route cases to response workflows
- Roll back the suspect revision and protect workload secrets
Explanation
Correct Answer
E:In the Cloud function unusual outbound connections case, trusted deployment and credential rotation stop a malicious function from recurring; the observation “A serverless workload begins contacting an unapproved Internet host after a dependency update” makes this response decision choice the fitting one.
Incorrect Answers
A:That option fits Google Cloud Security Command Center posture since a posture result requires workflow follow-through to reduce risk; it does not fit the Cloud function unusual outbound connections response decision scenario, where trusted deployment and credential rotation stop a malicious function from recurring.
B:This option addresses Cross-cloud incident correlation because containment must cover all active relying-party sessions, not just the IdP console; the Cloud function unusual outbound connections response decision question instead turns on the fact that trusted deployment and credential rotation stop a malicious function from recurring.
C:That option fits Google Cloud audit log principal investigation since response should address both stolen key paths and grants already made; it does not fit the Cloud function unusual outbound connections response decision scenario, where trusted deployment and credential rotation stop a malicious function from recurring.
D:This option addresses Google Security Operations threat correlation because actionable investigations depend on tuned content and clear ownership; the Cloud function unusual outbound connections response decision question instead turns on the fact that trusted deployment and credential rotation stop a malicious function from recurring.
Question 629
Management reports the number of alerts processed per analyst as proof the SOC detects attacks quickly. Which working hypothesis is best supported by these facts?
- Insufficient case evidence provenance
- Detection-content quality erosion
- Misleading volume-based SOC KPI
- Business-impact-aware incident prioritization
Explanation
Correct Answer
C:For classification in SOC management performance metric, throughput measures activity but not time to discover and contain genuine incidents; this applies to the reported “Management reports the number of alerts processed per analyst as proof the SOC detects attacks quickly”.
Incorrect Answers
A:In SOC case evidence preservation, incident records must retain context showing where an artifact came from and how it was interpreted; the reported SOC management performance metric case calls for a different classification conclusion because throughput measures activity but not time to discover and contain genuine incidents.
B:For the alternative SOC detection quality governance situation, high alert volume with low investigative value indicates a rule needs contextual tuning and measurement; by contrast, for SOC management performance metric classification, throughput measures activity but not time to discover and contain genuine incidents.
D:For the alternative Alert severity versus priority situation, priority combines technical confidence with asset criticality and possible operational harm rather than copying detector severity; by contrast, for SOC management performance metric classification, throughput measures activity but not time to discover and contain genuine incidents.
Question 630
An external threat feed description includes instructions telling a summarization assistant to delete the security case. What monitoring improvement would best help detect recurrence?
- Monitor precision, recall and feature stability over time
- Test enrichment prompts against hostile text samples
- Track false-positive automated actions and rollback success
- Detect secret patterns in outbound AI prompts where policy permits
Explanation
Correct Answer
B:The Prompt injection in threat enrichment evidence “An external threat feed description includes instructions telling a summarization assistant to delete the security case” is decisive for ongoing detection improvement because adversarial fixtures help identify instruction-following weaknesses in AI-assisted triage.
Incorrect Answers
A:For the alternative AI alert score drift situation, quality metrics are more informative than merely counting model alerts; by contrast, for Prompt injection in threat enrichment ongoing detection improvement, adversarial fixtures help identify instruction-following weaknesses in AI-assisted triage.
C:For the alternative Automated response approval boundary situation, operational metrics should measure disruption as well as time saved; by contrast, for Prompt injection in threat enrichment ongoing detection improvement, adversarial fixtures help identify instruction-following weaknesses in AI-assisted triage.
D:In Sensitive data entered into AI copilot, data controls help prevent accidental transmission from security consoles; the reported Prompt injection in threat enrichment case calls for a different ongoing detection improvement conclusion because adversarial fixtures help identify instruction-following weaknesses in AI-assisted triage.