Beginner·200 points·7 solves
Introduction to SOC Workflow
Learn the five-step SOC triage workflow on a small, realistic incident: read a SOC alert feed, pivot to the underlying auth log, establish a baseline for the account, and map the activity to MITRE ATT&CK. The technique is deliberately easy; the workflow is the lesson.
A SOC does not run on intuition — it runs on a repeatable loop: TRIAGE (is this alert worth a human?), INVESTIGATE (collect the evidence it points at), SCOPE (what is the account's normal behaviour, and how does this deviate?), ATTRIBUTE (which technique is this?), and DECIDE (how severe, and what next?). Two habits carry the whole loop. The first is that every claim needs a log line behind it — "the address looks foreign" is not a finding, "the address sits outside every range documented in the asset inventory" is. The second is that absence is evidence: what a log does not contain is often as informative as what it does, and that matters most when you come to attribute the activity. Evidence is in /var/log/soc: alerts.json (start here), auth.log, session-commands.log and asset-inventory.csv.
Your tasks
1
Triage the alert feed
Open /var/log/soc/alerts.json — one JSON object per line. Every alert carries a status and a severity. The platform auto-closes the patterns it already recognises as benign, but status alone will not isolate your incident: more than one alert is still open, and an open ticket is not automatically a security matter. Work out which single alert a Tier 1 analyst should pick up, and be ready to say why each of the others is not it.
2
Pivot to the evidence
The alert gives you a host and a time, not a person. Take both to /var/log/soc/auth.log and read the SSH authentication events on that host around that window: which account authenticated, from what address, and at exactly what second. auth.log omits the year — the alert feed supplies it, and answers are expected in UTC on a twenty-four hour clock.
3
Establish the baseline
A single login proves nothing on its own; "unusual" is a claim you have to support. Pull every line for that account out of auth.log and look at where it was connecting from earlier the same morning. Then test each source address against the approved corporate ranges listed at the foot of asset-inventory.csv, and note which host the inventory says that account owns.
4
Scope the session
session-commands.log tags every command with a session id, and it is ordered by time rather than grouped by user. Isolate the session belonging to the login you are investigating, count only its entries, and read them in sequence — the order tells you what the operator was trying to learn, and whether that resembles the account owner's ordinary work.
5
Attribute and decide
Map what you have to a single MITRE ATT&CK technique; the deciding detail is as much what the authentication log does not show as what it does. Then assign a severity from the standard scale and justify it out loud — weigh what the account could reach, whose data that host holds, whether the source was authorised, and how far the session got. Severity follows the evidence, not how alarming the incident feels.
Answer the questions
Start the lab, investigate the evidence in the terminal, and answer the questions here.