In the beginning …¶
The attacker starts somewhere. It might be an API endpoint with overly generous permissions, a web application that trusts input it really shouldn’t, a misconfigured cloud storage bucket, an endpoint with an unpatched kernel, a flat network that forgot segmentation existed, an OT controller accessible from the IT side, or a human who received a very convincing email at 16:00 on a Friday. This section covers defensive controls and detection for each of these initial access vectors — because the kill chain has to start somewhere, and the goal is to make “somewhere” as inhospitable as possible.