Skip to content
Tutorials Center
My learning

Topics

Lesson 37 of 37 ยท Projects and Practice

Practice: SOC Analyst Scenarios

Practice SOC decisions on six invented scenarios: triage, phishing, ransomware signs, insider data movement, alert tuning and metrics, with reasoning and answer keys.

Reading about incident response is not the same as deciding under pressure. This practice lesson gives you six short scenarios based on situations every SOC faces. Try to decide before you open each answer. All names, hosts and addresses are invented, and the "logs" are written for teaching; the scenarios involve no real systems.

For each scenario, answer three questions: What do I believe is happening? What do I check next? What do I do now, and who do I tell?

Scenario 1: the 3 a.m. administrator

An alert: administrator account admin-jlee signed in to the VPN at 03:12 local time from a country the user has never used. The same account then connected to a file server. The user's calendar (from the HR system) shows they are on leave.

Think like an analyst

Believe What probably happened?
Check What proves or disproves it?
Act What is safe to do now?

Decide first, then compare with the key.

1 / 2
Answer key

Belief: probable account takeover. Privileged account, unusual location and time, user on leave: several weak signals add up.

Check: authentication history for the account (failed attempts first? new MFA device? a password reset?), the VPN session's duration, what was accessed on the file server, other logins from that address. Call the user through a known number, not a number in the alert.

Act: this involves a privileged account, so escalate immediately. Revoke sessions and disable the account (reversible), then scope. Do not wait for the user to answer if you cannot reach them quickly. See triage and containment.

Scenario 2: the shared invoice

Three users report an email "Invoice 7741 attached". Two opened the attachment, and one entered credentials on a page that opened. The mail gateway marked the message "clean" because the link was added after delivery.

Answer key

Belief: phishing campaign with a delayed-weaponized link. "Clean" at delivery does not mean clean now.

Check: all recipients (search by sender, subject, link domain over a wide window), who clicked (proxy and DNS logs), who submitted data (POST requests), attachment sandbox result, endpoint alerts on the two who opened it.

Act: purge the message from all mailboxes, block sender and domain, reset the credentials and sessions of the user who typed them, scan the endpoints. Tell the three reporters. Follow the phishing playbook.

Scenario 3: files with a new extension

An endpoint tool shows one file server renaming thousands of documents in ten minutes, adding the extension .locked, and a text file named README-RECOVER.txt appears in each folder.

Answer key

Belief: ransomware encrypting files right now. This is a critical incident.

Check (quickly): which account and process are doing it, which shares are reachable from that server, whether other servers show the same pattern.

Act: immediately isolate the server from the network (do not power it off: memory may hold keys and evidence), disable the account in use, block the source host, alert the incident commander, start the IR plan, protect backups (disconnect them from the network), and decide communications with leadership and legal. Recovery is from clean images and verified backups only. Read the lifecycle and evidence handling.

Scenario 4: the large upload

A data-loss rule fires: an employee in the finance team uploaded 4 GB to a personal cloud storage site last night. The employee resigned last week.

Answer key

Belief: possible insider data exfiltration, but not proven. The person may be backing up personal photos.

Check: what the files were (names, sizes, classification, if logs allow), whether the destination is permitted by policy, previous behavior, any other signs (USB use, printing, emails to personal accounts).

Act: this is a sensitive HR and legal matter. Preserve evidence and keep the circle small: notify the incident manager, HR and legal before any confrontation or monitoring change. Do not accuse. Follow your organization's insider-threat process and privacy rules.

Scenario 5: the noisy rule

A rule named "Remote admin tool launched" produced 900 alerts last month; 880 were closed as false positives (IT staff using approved tools). Analysts want it disabled.

Answer key

Belief: the rule is noisy but protects against a real technique, so disabling it is risky.

Do: measure precision and recall, then add context: exclude approved tools used by members of the IT group on admin workstations, and keep alerting when the same tool is launched by a non-IT user or from an unusual parent process, or a tool that is not on the approved list. Back-test against past incidents. Give exclusions an owner and expiry. See alert tuning.

Scenario 6: the report to the board

Your manager asks for one slide: "How is the SOC doing this quarter?" Your data: mean time to acknowledge went from 12 to 25 minutes, the false positive rate fell from 80 to 55 percent, alerts rose 30 percent, and two analysts left.

Answer key

Reading: slower acknowledgment is explained by fewer analysts and more alerts, while quality improved because tuning worked. Present median and 90th percentile (not just means), state the staffing cause, show the quality gain, and propose a request (hire, automate enrichment, or further tuning). See SOC metrics.

Try it: triage drill

Practice ranking a queue. The scoring rules here are deliberately simple; change them and see how the order changes.

Python Rank a queue of scenario alerts

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Notice that activity in progress outweighs everything else: stopping an attack in progress is always the first priority, as in scenarios 2 and 3.

Try it: spot the false positive

Decide for each event whether to close or escalate, then see how a simple rule set agrees. The reasons matter more than the verdict.

Python Close or escalate with a written reason

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

How to keep practicing

  • Run tabletop exercises with colleagues: one reads a scenario, the others talk through each step.
  • Play capture-the-flag and blue-team labs on platforms and ranges that you are authorized to use.
  • Read public incident reports and post-mortems and write your own timeline and detections for them.
  • Replay your own organization's past incidents as tests for new detections.
  • Learn the data: know what each log source can and cannot tell you.

Practice only on systems and data you own or are explicitly authorized to test. Attacking real third-party systems is illegal, even "just to learn".

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. In scenario 3 (mass file renaming), why isolate the server from the network instead of powering it off?

    Memory can hold keys and artifacts; network isolation limits damage but keeps volatile data.
  2. 2. Scenario 4 involves a departing employee. What is the right early step?

    Insider cases are sensitive, so legal and HR processes must guide actions.
  3. 3. What is the best response to the noisy rule in scenario 5?

    Context keeps the detection while removing the approved use from the noise.
  4. 4. Which factor should most increase an alert's priority?

    Stopping an active attack limits damage the most, so activity in progress comes first.

Summary

  • Practice decisions with scenarios: believe, check, act, and say who you tell.
  • Activity in progress, privileged accounts and data risk raise priority; reasons for closing must be verifiable.
  • Insider cases need HR and legal; ransomware needs isolation and evidence preservation.
  • Tune rules with context rather than switching them off, and report metrics with their causes.

Next: SIEM and SOC Operations