root@thehacksparrow:~/writeups$ SYSTEM ONLINE
root@sparrow:~/writeups$ cat soc-l2-alert-triage.md
// Blue Team & SOC

SOC L2 Alert Triage

15 Mar 2023 · 11 min read
Overview — A room on the SOC Level 2 path centred on the core L2 responsibility: picking up an alert that L1 has already triaged and escalated, then performing the deeper log analysis, verdict-making, and advanced response actions that L1 cannot do alone. The idea running through the whole room is that L1 optimises for triage speed while L2 optimises for triage quality and depth. It closes with a hands-on challenge around a fake Claude Desktop installer that actually delivers an infostealer.
PlatformTryHackMe
CategoryBlue Team & SOC
DifficultyEasy
RoomSOC L2 Alert Triage

L1 versus L2

Before diving into the tasks it helps to pin down the role difference the room repeats throughout:

AspectLevel 1Level 2
TriggerNew security alertEscalated alert
FocusQuick alert triage within SLADeeper log analysis and response
ToolsTicketing system + SIEMWider range of SOC and IT tools
ResponseQuarantine a file, approve a SOAR playbookManually clean malware, disable users, isolate hosts

Task 1 — Introduction

No graded question.

This task sets the stage: an escalated alert may represent exfiltration, ransomware, or wiper activity. The room's objective is to teach the L2 workflow, SOC best practices, and the senior-analyst mindset.

Prerequisites:

  • Senior Security Analyst Intro room
  • SOC Team Internals module

Task 2 — SOC L2 Workflow

Q: What is the most common trigger for L2 to start the triage?
A: Escalated alert

Explanation

Unlike L1, whose workflow always begins with a fresh alert firing in the SIEM, an L2 analyst's day-to-day work begins when L1 has already reviewed an alert and pushed it upward because it needs deeper investigation, coordination with other teams, or an advanced response action. Occasionally an L2 may also be triggered by a direct urgent request from management or an MSSP customer, but escalation from L1 remains the dominant trigger.

Task 3 — Log Analysis as L2

Q: Should you understand the rule purpose before triaging the alert? (Yea/Nay)
A: Yea

Q: What term is used for a chronological list of events (related to the attack)?
A: Timeline

Explanation

Before touching any logs, the analyst must know exactly what technique the triggered detection rule is designed to catch — otherwise the investigation starts blind and risks missing the actual attack pattern.

Once the rule's purpose is understood, the L2 workflow follows two stages:

  1. See what happened — quick SIEM queries to build a high-level «story» (why the rule fired, what happened next, what caused the initial trigger).
  2. Build a detailed timeline — a full chronological reconstruction of process, file, and network activity across every involved host, used to surface hidden indicators such as dropped DLLs, C2 domains, or malicious IPs.

When the story alone cannot answer every open question, L2 analysis becomes a threat-hunting loop:

Form a hypothesis → build a timeline to close the gaps → evaluate whether the story is complete → repeat if it isn't.

Task 4 — Threat Response

Q: What is the term for a temporary response that stops the threat from spreading?
A: Containment

Q: You see clearly malicious activity on a corporate device. You also know about an ongoing pentest, but the red teamers do not respond. Would you isolate the device before receiving confirmation? (Yea/Nay)
A: Yea

Explanation

Response in the L2 workflow follows a strict logic gate:

Verify the activity -> Determine true/false positive -> Respond accordingly -> Resolve the case -> Capture lessons learned

Verifying the activity means confirming legitimacy directly with the affected party before jumping to conclusions: asking the user whether a login was theirs, confirming a new admin account with IT, or checking a suspicious API call with DevOps. If there is no response and the risk is high, the account or host gets disabled/isolated by default rather than left exposed.

Response to true positives scales with blast radius. A single infected laptop might only need an EDR quarantine, while a large-scale intrusion could require RDP/SSH access to hosts, disabling identities in Entra ID, subnet isolation, and emergency patching.

Response to major incidents follows the classic three-phase IR model:

PhaseAction
ContainmentIsolate hosts or disable users to stop the threat from spreading; applied immediately, even before the investigation is complete.
EradicationRemove malware, rotate stolen credentials, revoke exposed privileges.
RecoveryLift containment, patch the root-cause vulnerability, and monitor for reinfection.

This is why isolating a device with clearly malicious activity is the correct call even when a pentest is technically in progress — without confirmation from the red team you cannot assume the activity is sanctioned, and the cost of a false containment is far lower than the cost of an active real intrusion left unchecked.

Response to false positives still requires action, just at lower urgency — three common outcomes:

  • False Positive Tuning — fixing a flawed detection query.
  • Security Hardening — fixing an exposed risk the FP incidentally revealed (e.g. open RDP).
  • Team Improvement — correcting an L1 analysis mistake through mentorship.

Resolving the case always includes three non-negotiable steps:

  1. Log evidence of every action taken in a centralised ticketing system.
  2. Inform the wider SOC team (especially about anything novel).
  3. Keep external parties such as MSSP customers updated with a summary.

Task 5 — Learning Lessons & Practical Challenge

This task combines a reflection principle with a hands-on simulation around a fake Claude Desktop installer that actually delivers an infostealer.

Challenge Phase 1 — Analysis (drag-and-drop into buckets)

See what happened (first 5 minutes)

  • Quickly review what the malware is doing now and how urgent the situation is.
  • Check the past activity: who launched Claude and where it was downloaded from.

Build a timeline (detailed investigation)

  • Trace and note the malware's full process, file, and network activity up to the present time.
  • Reconstruct the complete attack timeline and collect indicators of compromise (IOCs).

Make your verdict (summarise what you've learned)

  • Reach a True Positive verdict and proceed to the Response stage with the collected IOCs.
  • Determine the root cause of Claude running malware: supply-chain compromise, prompt injection, or an imposter binary.

Challenge Phase 2 — Response Actions (ordered sequence)

The scenario reveals that the «Claude Desktop» installer was actually an infostealer downloaded from a fake claude.top website. The correctly ordered response sequence:

  1. Isolate the infected LPT-1601 from the network (Containment).
  2. Eradicate the infostealer and its files from the host (Eradication).
  3. Rotate all stolen passwords, keys, and tokens (Eradication).
  4. Lift host isolation and inform the user and your team (Recovery).
  5. Monitor — ask your team to keep an eye on the user/host (Recovery / monitoring).
This directly mirrors the NIST Containment → Eradication → Recovery model taught in Task 4: isolate first to stop spread, clean the host and rotate compromised credentials, then only lift isolation once the threat is fully removed, followed by continued monitoring for reinfection.

Challenge Phase 3 — Lessons Learned (select the 4 best options)

The four correct reflection actions for this incident:

  1. Search for attack indicators on the other workstations.
  2. Communicate with your L1 on how to improve escalation comments.
  3. Build an early-stage detection rule that covers malicious downloads.
  4. Suggest enabling a web filter to block fake websites.

Incorrect distractors: replacing the EDR vendor outright and disciplining the affected user. Both go against the senior mindset this room teaches: a single miss doesn't justify a vendor swap, and blaming the end user contradicts the «learn, don't punish» culture senior analysts are expected to build within their SOC team.

Flag

The challenge flag is generated dynamically per session inside the embedded static site once all three phases are completed correctly, so it will be unique to your run and appear directly on the results screen after phase 3 is checked.

Task 6 — Conclusion

No graded question.

This task confirms room completion and points toward the next room, Report Writing for SOC L2, which builds on this incident by teaching how to document it in a professional report for stakeholders and MSSP customers.

Key takeaways

  • L2 triage begins with escalation, not raw alerts — always understand the detection rule's purpose first.
  • Investigation follows «see what happened» → «build a timeline», looping as a threat hunt when the story is incomplete.
  • Response follows a strict logic gate: verify → classify → respond → resolve → reflect.
  • Major incidents follow the Containment → Eradication → Recovery model, with containment applied immediately even before investigation completes.
  • Case resolution always requires logging, internal communication, and external updates.
  • Lessons learned should focus on systemic improvement (detection rules, escalation quality, hardening) — not vendor blame or user punishment.