Question 160
A WAF blocks hundreds of suspicious requests to an application but the backend audit logs show no matching execution. How should triage characterize the evidence?
- Evidence that all application vulnerabilities have been remediated
- Likely blocked attack attempts, subject to outcome validation
- Certain successful compromise of the backend database
- Proof that the request source is benign because it was denied
Explanation
Correct Answer
B:Blocked requests are security events; confirmed compromise requires evidence that a harmful request reached or altered the backend.
Incorrect Answers
A:A functioning filter can mitigate observed requests but does not by itself fix unsafe application code or every bypass.
C:A WAF detection cannot demonstrate execution when the application records show no corresponding effect.
D:Blocking does not establish intent or innocence; the request pattern may still represent an active adversary.
Question 161
A Windows endpoint records an unfamiliar Run key addition associated with a program in a user-writable folder. Which analytic is appropriate?
- Filter exclusively on DNS query lengths
- Join autorun changes with the launched file and parent process
- Monitor only externally opened firewall ports
- Treat every registry change as confirmed malicious persistence
Explanation
Correct Answer
B:Startup persistence is stronger when the registry change points to an unexpected executable and an identifiable actor.
Incorrect Answers
A:DNS features do not directly observe modification of Windows autorun configuration.
C:Local registry persistence may exist even with no inbound listening service or external port exposure.
D:Administrative software routinely updates registry keys; the key path and executable context are what matter.
Question 162
An EDR flags an unsigned binary spawned from a trusted system utility; the SIEM shows its network connections to a rare destination. What combined conclusion is appropriate?
- Discard the EDR alert because network logs are already available
- Declare all binaries launched by signed Windows utilities harmless
- Assume a rare destination means data theft has definitely completed
- Investigate the host process alongside the suspicious network flow
Explanation
Correct Answer
D:Process ancestry and outbound traffic jointly raise concern, while file provenance and authorized use still require verification.
Incorrect Answers
A:The EDR supplies process and file context that network metadata alone cannot reconstruct.
B:Trusted tools can be abused as parent or execution mechanisms, so signer status does not guarantee benign child activity.
C:Rarity supports investigation but does not prove content or volume of an exfiltration transfer.
Question 163
A cloud administrator disables a security logging service and immediately creates a new access key for an unfamiliar identity. Which rule structure fits?
- Correlate audit-policy changes and new credentials by actor and time
- Assume security logging cannot be disabled in cloud environments
- Monitor only the CPU consumption of virtual machines
- Alert on every new key without considering approval or preceding logging changes
Explanation
Correct Answer
A:Combining control-plane operations reveals potential defense evasion and persistence that neither event alone conclusively establishes.
Incorrect Answers
B:Cloud audit controls and service integrations can have changeable configuration, making control-plane monitoring important.
C:Control-plane changes can occur without any VM CPU anomaly, so compute metrics do not cover the targeted actions.
D:Key creation can be normal; sequence and change authorization improve confidence and prioritization.
Question 164
A SIEM runs a rule every five minutes against the last five minutes of data, but event ingestion can lag ten minutes. What risk should the engineer address?
- The SIEM will automatically correct every late event without any configuration
- Late arrivals may be missed unless the rule uses an overlap or delayed-event strategy
- Converting every source timestamp to its ingest time ensures original chronology
- Running the query less often always increases accuracy
Explanation
Correct Answer
B:Query cadence must account for ingestion latency, deduplication, and event time to avoid permanent coverage gaps.
Incorrect Answers
A:Not every scheduler rechecks historic windows; assuming automatic replay can leave silent detection gaps.
C:Ingest time can distort event order and masks the very delay that the analytic needs to handle.
D:Reduced frequency may save compute but does not inherently repair missing late-arriving data.
Question 165
An alert says a domain controller contacted a newly malicious IP but provides no process context. What should the SOC analyst do before declaring compromise?
- Close the alert because domain controllers sometimes make network connections
- Check service role and flow context against current intelligence
- Assume the domain controller initiated the flow because its name appears in the title
- Immediately publish attribution to a named intrusion group
Explanation
Correct Answer
B:A single address match is an indicator; host and traffic context determine whether the observation reflects malicious communication or a false positive.
Incorrect Answers
A:Normal connectivity does not excuse traffic to a suspicious destination; the specific service and timing still require analysis.
C:Alert labels can contain normalized or translated addresses; the underlying event must confirm direction and actor.
D:A reputation match alone cannot establish actor identity, motive, or a complete intrusion narrative.
Question 166
A confirmed compromised domain-admin account is actively issuing directory changes. Which factor most strongly supports critical incident severity?
- The incident ticket is the oldest one in the queue
- Prioritize confirmed high-privilege malicious activity
- The finding contains only four distinct tags
- The log source is popular among SOC analysts
Explanation
Correct Answer
B:Severity should reflect confidence and consequences, not merely alert count; domain-wide administration increases blast radius.
Incorrect Answers
A:Age matters to service management, but a confirmed privileged intrusion is urgent because of active impact and access.
C:Tag count is not a credible proxy for confirmed malicious actions or affected assets.
D:Familiarity with a product does not measure adversary capability or exposure of identity infrastructure.
Question 167
A rule warns when a signed software deployment service starts a shell. Repeated alerts match a documented rollout on approved hosts. What is the safest tuning action?
- Disable audit process creation for the organization
- Scope an exclusion to verified deployment identity, host group, parentage, and change window
- Mark all past and future alerts from the deployment service as benign without testing
- Suppress every command-shell event on every host
Explanation
Correct Answer
B:A narrow documented condition reduces routine noise while keeping anomalous shells from that service visible outside approved contexts.
Incorrect Answers
A:Removing telemetry impairs many unrelated detections and prevents validation of future suspicious execution.
C:A known benign pattern can be impersonated; changes in command, target, or execution context still matter.
D:Attackers frequently use shells, so a global exclusion would create a large blind spot.
Question 168
Two detection variants fire on the same test dataset: rule A catches 18 of 20 attacks with 40 false alerts; rule B catches 17 with 2 false alerts. What should reviewers consider?
- Declare both rules equivalent because their true-positive counts are close
- Both detection recall and false-positive workload against business risk
- Choose solely the rule that emits the most alerts
- Choose solely the lowest compute-cost rule without examining missed attacks
Explanation
Correct Answer
B:Rule B greatly reduces triage noise but misses one extra test case; choosing requires understanding which attack was missed and its impact.
Incorrect Answers
A:The false-positive difference is substantial and affects analyst capacity and trust in the detection.
C:A high count can reflect noise rather than effective identification of malicious behavior.
D:Cheap execution is not valuable if it fails critical detection objectives or creates unacceptable exposure.
Question 169
An analyst sees a suspicious process hash in an alert. What evidence should be checked before assigning a malware family?
- Only whether the process name uses capital letters
- Only the alert title written by a third-party feed
- Only the local username length
- Corroborate file provenance with observed execution behavior
Explanation
Correct Answer
D:A digest or suspicious filename is not enough for confident family attribution, especially when samples may be modified or misclassified.
Incorrect Answers
A:File-name typography has little evidentiary value for malware classification and can easily be mimicked.
B:Vendor labels can be mistaken and do not replace evidence of what the executable did on the affected system.
C:Account string length does not identify a binary's capabilities or origin.
Question 170
An escalated incident case needs to remain actionable when ownership changes. What content should its analyst note preserve?
- Only an unstructured sentence saying the alert looks suspicious
- Only the final status without rationale or artifacts
- Record the investigation timeline, evidence, and open decisions
- Only the analyst's personal opinion of the detection vendor
Explanation
Correct Answer
C:Structured case notes permit reproducible reasoning and show the next responder what was actually observed versus inferred.
Incorrect Answers
A:A vague assessment omits supporting events, impacted systems, and the exact next decision.
B:A status field cannot explain why a case was escalated or how another analyst should validate it.
D:A vendor impression does not provide evidence or a defensible incident timeline.