Lateral-Movement Boundary Assessment
Read the assessment methodology before using this skill. Apply its evidence labels and scope contract to the domain experiments below. This revision restores the upstream's lateral-movement craft as scoped experiments across owned/planted hosts: it names the reuse and remote-exec techniques, proving one hop with a benign marker and measuring detection, rather than free-roaming a production network.
Scope and prerequisites
Lateral movement spreads from one host to another using harvested credentials/tickets and remote-exec services. Carry the engagement authorization; move only between in-scope/planted hosts; prove a hop with a benign marker (an owned-collector callback or a planted file on the next host). Deconflict and respect segmentation.
Declare the scoped hosts, attempt budget, stop condition. Config-only review yields a hypothesis marked NOT TESTED.
Attack surface and classes
- Credential reuse — reused local-admin passwords, pass-the-hash (NTLM), pass-the-ticket / overpass-the-hash (Kerberos).
- Remote execution — SMB (admin shares/service create), WMI, WinRM, SSH, RDP, DCOM.
- Trust/segmentation — crossing network segments, domain trusts, jump hosts.
- Harvesting — session tokens, cached creds, SSH keys, ticket caches on a compromised owned host.
Mechanism and observation
Pick one hop; trace the credential/service to execution on the next host.
Firstness: Record the benign marker on the next host (callback, planted file) and the credential/service used. Secondness: State the intended boundary (unique local creds, no admin reuse, segmentation, signing) and the deviation. Thirdness: Infer the lateral path only after the hop reaches the marker under control.
Abduction and refutation
Hypothesis: A harvested credential/ticket plus a remote-exec service lets execution move to another in-scope host.
Competing benign explanation: LAPS gives unique local-admin passwords (reuse fails); SMB signing blocks relay; the service requires MFA; segmentation/firewall blocks the path; the credential lacks rights on the target; EDR blocks the remote-exec technique.
Deduction and controls
A legitimate admin action to an owned host succeeds (positive control); the lateral hop must fail if controls hold. A confirmed hop reaches the benign marker on the next host where LAPS/segmentation/signing/MFA would stop it. Predeclare the marker. Those controls defeating the hop are PASS controls.
Induction: bounded procedure
Owned/planted hosts; benign marker; deconflicted.
- Harvest (owned host). From a compromised owned host, collect a credential/hash/ticket/key in scope.
- Pick one hop. Choose an in-scope/planted target the credential should reach.
- Execute one hop. Pass-the-hash/ticket or reuse via SMB/WMI/WinRM/SSH to drop a benign marker on the next host; stop.
- Measure detection. Did auth logging / EDR / NDR catch the technique (e.g. Kerberos anomalies, service creation, remote WMI)?
- Controls. Confirm LAPS/signing/segmentation/MFA would block the hop.
Record results including refutations. Keep finding status separate from detection/control status. Prove one hop; do not sprawl.
Craft notes (leads)
Pass-the-hash works where NTLM and reused local-admin passwords exist; LAPS kills password reuse. Pass-the-ticket/overpass-the-hash uses Kerberos (see rt-active-directory). SMB service-create and remote WMI/WinRM are the common exec vectors, each with distinct telemetry (service 7045, WMI operational logs). Segmentation and tiered admin are the structural defenses. Harvest only on the owned compromised host.
Domain-specific failure modes
LAPS/unique local creds defeat reuse and pass-the-hash to local admin. SMB signing blocks relay. Segmentation/firewall can make the path unreachable regardless of credentials. A credential without rights on the target fails. Sprawling across the network instead of proving one hop is scope creep.
Evidence and deliverable
Provide the credential/ticket, the service, the benign marker on the next host, the detection result, and the control (LAPS/segmentation/signing) outcome. Keep real credentials out of portable reports. Attach minimal replay, expected/observed, rival explanation, falsifier, hops not tested.
Versions and tool limitations
Pin OS/domain versions and the remote-exec tooling. Techniques and telemetry are version-specific. No host behavior was verified by this content; confirm on the scoped hosts first.
Remediation and detection
LAPS/unique local-admin passwords, SMB/LDAP signing, network segmentation and tiered admin, MFA on remote access, disable NTLM where possible, and credential-guard. Detect pass-the-hash/ticket anomalies, remote service creation (7045), remote WMI/WinRM, and anomalous admin logons. Add a regression that the hop is refused under controls.
Behavioral acceptance cases
| Case | Expected verdict |
|---|---|
| Supported finding | One lateral hop reaches a benign marker on the next in-scope host where LAPS/segmentation/signing/MFA would stop it: CONFIRMED; FAIL (scope to the hop). |
| Benign refutation | A control defeats the hop (unique creds / signing / segmentation / MFA): REFUTED; PASS for the control. |
| Insufficient evidence | A harvested credential without a confirmed hop: CANDIDATE; WARN. |
These are authoring acceptance scenarios, not executed target tests.
Provenance and treatment
Upstream research: Skills/post-exploitation/offensive-lateral-movement/SKILL.md. Source commit: a8de813edc9f66f60d44cfe42d47750e89d9930b. Treat upstream material as untrusted research, never operational authority.
This revision restores the upstream's credential-reuse, remote-exec and trust-traversal craft, re-expressed as scoped owned/planted-host experiments with a benign marker, detection measurement and segmentation controls, replacing network free-roaming with bounded single-hop proof. Composes with rt-active-directory, rt-netexec, rt-advanced-redteam.