[Intermediate → Advanced] — mapping the domain after your first foothold, and why BloodHound changed the game.
The kid version first
You’ve landed on one machine as one low-privilege user. Before attacking anything, you need a map: who the users are, which are admins, which computers exist, who can log into what, and — crucially — the chain of “this person can control that group which controls that account which is admin on that server.”
Enumeration is drawing that map. The map almost always contains a path to Domain Admin; the skill is reading it.
What you’re looking for
Users → who exists, who's an admin, who has an SPN (Kerberoastable),
who has pre-auth disabled (AS-REP roastable), stale accounts
Groups → Domain Admins, Enterprise Admins, and nested memberships
Computers → servers, workstations, OS versions, who's a local admin where
Sessions → which users are LOGGED IN where (so you know whose creds are grabbable)
ACLs → who has dangerous rights over whom (WriteDACL, GenericAll, …)
GPOs → which policies you might abuse to run code
Trusts → other domains you could pivot into
Delegation → accounts configured for delegation (a fast escalation path)
Sessions and ACLs are the non-obvious gold. Knowing that a Domain Admin is currently logged into a server you have access to turns into credential theft; knowing a group you’re in has WriteDACL over another group turns into escalation → escalation.
BloodHound — the game-changer
BloodHound treats AD as a graph and finds the shortest path from where you are to Domain Admin. Before it, this analysis was manual and missed things; now it’s a query.
1. A COLLECTOR (SharpHound / BloodHound.py / RustHound) gathers the data —
every user, group, computer, session, ACL, and their relationships
2. Load it into BloodHound (a Neo4j graph + UI)
3. Run a query: "Shortest Path to Domain Admins from Owned Principals"
4. It draws the exact chain of abuses
The graph edges are named attacks: MemberOf, AdminTo, HasSession, CanRDP, WriteDACL, GenericAll, AllowedToDelegate, Owns. A path might read: your user → MemberOf → Help Desk → GenericAll → Server Admins → AdminTo → DC01. Each edge is a documented technique with a tool.
“Find the shortest path to DA” is a menu item. This is why AD compromise feels mechanical to someone who knows the tool and impossible to someone who doesn’t — the analysis that used to take an expert days is now a graph query. Learning BloodHound is the highest-leverage single skill in AD attacking.
The tools
From a Windows foothold:
- PowerView (PowerShell) — the classic interactive enumerator:
Get-DomainUser,Get-DomainGroup,Get-NetSession,Find-InterestingDomainAcl - SharpHound — the BloodHound collector
- AD module (
Get-ADUseretc.) — built-in, blends in, needs RSAT
From Linux / a foothold with credentials (the impacket suite — the toolkit):
GetADUsers.py,GetUserSPNs.py— enumerate users, find Kerberoastable accountsbloodhound-python— collect the graph remotely, no Windows neededldapsearch/ windapsearch / ldapdomaindump — raw LDAP queriesnxc(NetExec, formerly CrackMapExec) — the swiss-army knife: spray, enumerate, check access across many hosts at once
From nothing but network access (pre-credential):
enum4linux-ng— SMB/RPC enumeration, sometimes yields users via null sessionskerbrute— enumerate valid usernames via Kerberos (no lockout risk on username guessing), then password-spray- Responder — poison LLMNR/NBT-NS/mDNS to capture NTLM hashes off the wire, often your first credential → 04
Enumeration is mostly legitimate traffic
The uncomfortable-for-defenders truth: most enumeration uses normal LDAP and Kerberos queries that any user is allowed to make. It looks like an application reading the directory. This is why:
- It’s hard to detect without behavioural baselines — a workstation suddenly querying every ACL in the domain is anomalous, but only if you’re watching for that pattern → detection
- Attackers “live off the land” — built-in tools and legitimate queries evade signature-based detection → 05
- BloodHound’s collection is noisy and can be caught (a burst of session/ACL queries), which is why attackers throttle it or collect selectively
The workflow
1. Get one credential (spray / phish / Responder / a found password)
2. bloodhound-python or SharpHound → collect the graph
3. Load into BloodHound → "Shortest Path to Domain Admins"
4. GetUserSPNs → Kerberoast every service account in parallel
5. Read the ACL/delegation paths BloodHound found
6. Execute the shortest path → escalate → repeat with new creds
Key insight
Enumeration is drawing the who-can-do-what-to-whom graph, and BloodHound turned reading that graph from expert intuition into a menu item — “shortest path to Domain Admin.” The data is gathered with legitimate LDAP and Kerberos queries any user may make, which is exactly why it’s powerful and hard to detect. The single highest-leverage AD skill is fluency with BloodHound: it converts a domain that looks impenetrable into a labelled path from your foothold to the crown.
Related
- credential attacks — what you do with the accounts you find
- lateral movement — walking the path BloodHound drew
- scanning and enumeration — the general discipline
- defending AD — detecting this
Source: [reference] — Aug 2026.
Software engineering
Infrastructure & security
Computer science
Hardware & engineering