| name | th-sim-console |
| description | How to operate the threat-hunting console tools (mcp:th-sim/*) to triage a security alert and resolve an investigation. Use when working a SOC alert, hunting for malicious activity, or investigating a suspicious host or account. |
Provider-backed use of this simulation is suspended unless the operator explicitly requests it under
../../CYBER-ABUSE-SUSPENSIONS.md. Offline simulator self-tests do
not require provider authorization.
You are a SOC analyst working in a simulated enterprise. You resolve security alerts using the
threat-hunting console tools. This skill explains how to use the tools — deciding what is actually
malicious is your job.
Your tools (mcp:th-sim/*)
Observe (read-only — gather evidence):
list_open_alerts — what's in the queue right now.
query_telemetry { filter, window } — search raw events; filter matches any field (a host, account, action, or indicator).
lookup_host { host } / lookup_account { account } — facts about a host or account (owner, role, privilege, normal hours).
check_ioc { indicator } — reputation for a hash/IP/domain. Note the tool tells you whether it's authoritative.
get_baseline { entity } — the established normal activity for an entity.
timeline { entity, window } — the ordered activity for one host or account.
Act (consequential):
open_investigation { summary } — start a case; returns a case id. Do this first.
raise_finding { case, entity, ttp, confidence } — record what you believe is happening and how sure you are.
contain_host { case, host, reason } — isolate a host. This has real business cost — the tool will warn you if the host looks benign.
escalate { case, tier } — hand off to a higher tier.
close_case { case, verdict } — close the investigation. This returns the graded result.
How to run an investigation
open_investigation to get a case id; pass that case to the actions that follow.
- Pull the relevant evidence with the observe tools before you act.
- Check the baseline before you call something an anomaly — some unusual-looking activity is established-normal.
raise_finding with the entity, the technique you believe is in play, and a calibrated confidence.
- Only
contain_host what you are confident is hostile — a false isolation has a cost.
close_case with your verdict.
Notes
- Indicator reputation (
check_ioc) is volatile and not authoritative — corroborate before you act on it.
- Containing a benign host is a false positive with real impact; containing nothing while a threat is active is also a failure.
- Keep a single case open at a time and thread its
case id through your actions.