Question 200
An AI alert summarizer asserts that a cloud identity is compromised, but the cited sign-in events are routine. What should the analyst do?
- Verify AI summaries against raw evidence and correct unsupported claims
- Automatically accept the summary if it includes an ATT&CK label
- Disable the identity solely because the summary uses confident wording
- Delete the raw identity logs to avoid conflicting conclusions
Explanation
Correct Answer
A:Generated interpretations can be mistaken; security decisions should rest on reproducible log observations and documented context.
Incorrect Answers
B:Framework labels can also be misapplied and do not prove that the underlying event occurred or was malicious.
C:Confidence of phrasing is not evidence of malicious activity and can cause disruptive false positives.
D:Preserving source records allows validation and later review; deleting them would destroy the basis for assessment.
Question 201
A SOC wants to send full incident logs containing passwords and personal data to a third-party AI service for summarization. What is the best first control?
- Apply approved data handling and redaction before external processing
- Disable evidence access logging so the AI job runs faster
- Send all fields because summary accuracy overrides data security
- Assume any AI endpoint is within the organization's security boundary
Explanation
Correct Answer
A:Incident logs can include secrets and regulated data, so authorization and privacy safeguards must precede external processing.
Incorrect Answers
B:Eliminating access records weakens accountability for sensitive case data and does not solve privacy obligations.
C:Unrestricted transfer may breach policy and expose credentials, even if the generated narrative is more detailed.
D:Service location and contractual protections cannot be inferred from the presence of an AI interface.
Question 202
Anomaly detection ranks an employee as high risk because login times shifted after a documented night-shift assignment. What should improve the model?
- Treat model output as conclusive proof of malicious insider activity
- Revalidate changed user context before recalibrating anomalies
- Delete the historical identity data used to understand prior behavior
- Permanently suppress every future alert about that employee
Explanation
Correct Answer
B:Behavior models need changing business baselines so legitimate shifts do not overwhelm analysts while meaningful deviations remain detectable.
Incorrect Answers
A:Anomaly scores indicate deviation, not malicious intent; schedule context can fully explain the change.
C:Removing all history discards baseline information and cannot explain the policy-driven schedule change.
D:One benign change does not eliminate future compromise risk or justify complete monitoring exclusion.
Question 203
An investigator using Splunk needs to narrow a search to a specific index and last 30 minutes before computing a count by source. Which approach helps performance?
- Disable parsed fields such as src_ip before using a by-source aggregation
- Convert the search to a dashboard screenshot and stop collecting logs
- Search every accessible index over all historical time and filter only after aggregation
- Apply index and time constraints early, then aggregate the resulting relevant events
Explanation
Correct Answer
D:Selective base searches minimize scanned data before expensive grouping, producing faster targeted incident analysis.
Incorrect Answers
A:The grouping depends on a stable source attribute, so removing it prevents meaningful aggregation.
B:A static image cannot run field-based searches or investigate new events as they arrive.
C:Global scanning is costly and introduces unrelated records into results before the useful restrictions are applied.
Question 204
An Elastic-based SOC detects that one data stream's documents stopped arriving after an integration update. What should engineers investigate?
- Assume the monitored threats have stopped because the index is empty
- Ingest pipeline, agent status, field-mapping errors, and data-stream health
- Only the age of the analyst's browser installation
- Only whether the Kibana dashboard title contains the correct company name
Explanation
Correct Answer
B:Connector or pipeline changes can interrupt normalization or indexing; end-to-end checks locate the gap in collected documents.
Incorrect Answers
A:A silent data source cannot establish that malicious activity ended; ingestion health needs verification.
C:A local browser does not explain missing documents in backend ingest pipelines.
D:A display title does not control agent collection or index insertion into a data stream.
Question 205
A SIEM detects an identity using a new device, new application, and unusual download volume. Which UEBA principle makes the combination useful?
- Assume all changes from baseline are malicious regardless of authorized business shifts
- Replace all access controls with a behavior score
- Rely only on the enterprise-wide average number of employees
- Evaluate deviations from the identity and peer-group baseline together
Explanation
Correct Answer
D:User and entity behavior analytics can find meaningful combined anomalies that isolated thresholds might treat as minor.
Incorrect Answers
A:New roles, devices, and travel can be legitimate, so context and validation remain essential.
B:Risk scoring supports detection and prioritization; it is not a complete substitute for authentication and authorization.
C:An organization-level count provides little information about the specific user's behavior.
Question 206
A SOC wants a dashboard showing affected business services rather than only hostnames. What must feed the visualization?
- Only the total number of successful user sign-ins per day
- Current service dependencies, asset owners and criticality
- Only general internet threat-news headlines
- Only a raw list of IP addresses without application context
Explanation
Correct Answer
B:Mapping infrastructure events to business services turns low-level detections into a view of possible operational impact.
Incorrect Answers
A:Login totals do not reveal which applications depend on the potentially compromised hosts.
C:External trends may inform awareness but do not connect this organization's assets to specific supported services.
D:IP inventories cannot identify which customer-facing service is at risk without dependency mapping.
Question 207
The SOC publishes a company-wide dashboard containing full usernames, private email addresses, and incident notes. What should change?
- Assume dashboard read-only permissions remove confidentiality risk
- Add more screenshots to the public dashboard without reviewing access
- Remove all access logs for dashboard viewers
- Apply role-based report access and publish masked summaries
Explanation
Correct Answer
D:Operational investigations need detailed identity records, but general viewers rarely need personal case data.
Incorrect Answers
A:A read-only user can still view or copy exposed personal and incident information.
B:Additional presentation does not address overexposure of sensitive identities and incident details.
C:Access auditability is important for sensitive data; deleting it weakens accountability and compliance.
Question 208
An auditor asks whether security events for privileged account changes were retained for the required period. Which report is most relevant?
- Document coverage and retention of required log sources
- A list of firewall brands used in branch offices
- A chart of total SOC employee attendance at training sessions
- A report of how many dashboards have been created
Explanation
Correct Answer
A:Compliance evidence must demonstrate that required event types were captured and preserved, not merely that a SIEM license exists.
Incorrect Answers
B:Network inventory alone does not establish which identity events were captured and retained for the audited duration.
C:Staff training records may support other controls but cannot verify the retention of privileged-account audit events.
D:Dashboard count is not proof that mandated security records remain searchable or preserved.
Question 209
A manager wants to determine whether incident containment is improving quarter over quarter. What is the strongest comparison?
- A count of all alerts regardless of confirmed status
- Total SIEM license costs without incident timestamps
- Number of raw network connections observed each quarter
- Percentiles of time to contain for equivalent incidents
Explanation
Correct Answer
D:Comparable incident classes and timestamps reduce distortion from outliers or changes in classification and case volume.
Incorrect Answers
A:Unfiltered detections reflect sensor and rule changes as well as threats and do not directly measure containment time.
B:Spending does not quantify elapsed time between confirmation and effective containment.
C:Traffic volume varies with workload and cannot establish how quickly responders contained incidents.
Question 210
A SIEM integrates with a SOAR platform that can quarantine emails after analyst approval. What is the primary architectural separation?
- The SIEM must automatically carry out every containment action without authorization
- The SOAR replaces all event sources so the SIEM no longer needs logs
- The email gateway alone determines incident severity for the enterprise
- Use SIEM for detection and SOAR for approved action orchestration
Explanation
Correct Answer
D:A response orchestration engine applies workflows and integrations, while the SIEM remains a central analytics and evidence platform.
Incorrect Answers
A:Detection and response can be separate, and approval gates are important for actions with business impact.
B:Orchestration requires evidence or alerts to act upon and does not magically reproduce collected security telemetry.
C:A downstream enforcement tool lacks the whole incident and business context needed for cross-system prioritization.