Commands
What to run, indexed by the question — and the commands that answer a different one.
What the course covers
- What is running on this machine right now? — The first thing anyone does, and the first place people over-read. A live process list is the state of the machine at the moment you looked, which is after the incident and after your own arrival.
- What ran before, and when? — The question the whole execution-artifact half of the corpus exists to answer, and the one with the most artifacts that mean something slightly different from what people assume.
- What will run again after a reboot? — Containment fails on what you did not enumerate. There are dozens of autostart locations and every built-in tool covers a different subset of them.
- Who authenticated, from where, and how? — Attribution in every engagement comes down to reading a logon record precisely — what it attributes, and what it cannot.
- What did this machine talk to? — The pivot that turns one host into a scope, and the artifact class with the shortest retention.
- What was written, moved or deleted? — Filesystem metadata is the densest timeline source on a Windows host and the easiest to misread, because there are eight timestamps and they are updated by different things.
- What does the logging actually cover? — Every conclusion rests on this and almost nobody checks it first. An absence means nothing until you know whether the source was recording.
- What is this file, and where did it come from? — Identification, provenance and whether it is what it claims to be — asked of every artifact anybody hands you.
- How do I collect this without ruining it? — Order of volatility, and the fact that your own commands are evidence too. The commands in this group are the ones most likely to be run badly under pressure.