Back to skills
SKILL.md
Windows Ad
ASecurityUse for authorized Active Directory and Windows identity attacks including Kerberos, AD CS, BloodHound paths, NTLM relay, and domain privilege escalation research.
- 2 stars
- 0 votes
- 0 copies
- 0 views
- Added October 17, 2026
Security analysis
100/100Pro scans all 3 files and shows the line behind each finding
npx -y skills add NetVar1337/agent-skill-canon --skill windows-ad --agent claude-codeAre you the author of Windows Ad?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/netvar1337-windows-ad)---
name: router-reverse-skill-router-windows-ad
description: Use for authorized Active Directory and Windows identity attacks including Kerberos, AD CS, BloodHound paths, NTLM relay, and domain privilege escalation research.
---
# Windows / Active Directory Security
## ACTION REQUIRED (execute immediately after reading)
1. `NOW`: read `../field-journal/precedent-pentest.md`
2. `NOW`: **domain/AD testing requires a clearly authorized scope** (including DCs, and whether poisoning/relaying is allowed)
3. `NOW`: case-init; write the network_profile and forbidden actions clearly
4. `NEXT`: tool-index (impacket/certipy/bloodhound etc. are often manual)
5. `ACT`: start from identity enumeration and the BloodHound graph, not destructive exploitation
## Applicable Scenarios
- Domain penetration, Kerberoasting, AS-REP, delegation
- AD CS (ESC1–ESC8 etc.) certificate attacks
- BloodHound / SharpHound attack paths
- NTLM Relay / Coercer forced authentication
- Local privilege escalation to domain paths (Potato etc. as a pivot)
## Relationship with attack-chain
- **Multi-stage from internet to domain controller** → PRIMARY can remain `attack-chain/`, this skill is the **AD specialist**
- **Already inside the domain, focused on identity** → PRIMARY = this skill
## Workflow
### 1. Enumeration
```bash
# Impacket examples / built-ins (require credentials and authorization)
nxc smb <range> -u user -p pass
bloodhound-python -d domain.local -u user -p pass -c All -ns <DC>
```
### 2. Common Paths (map first, shoot later)
```text
□ Kerberoast / AS-REP → offline cracking
□ ACL abuse (GenericAll/WriteDacl)
□ Delegation (unconstrained/constrained/resource-based)
□ AD CS template misconfiguration → Certipy
□ Relay: LLMNR/NBT-NS + ntlmrelayx (confirm authorization)
```
### 3. Credentials and Lateral Movement
```text
□ secretsdump / lsassy / mimikatz (strict authorization and cleanup)
□ PtH / PtT / golden tickets only within the authorized red-team scope
□ Write Evidence at every step; wait for user confirmation on high-risk actions
```
## Toolchain
| Tool | Purpose |
|------|------|
| BloodHound / SharpHound | Path graph |
| Certipy | AD CS |
| Impacket / NetExec | Lateral movement and enumeration |
| Rubeus / Mimikatz | Tickets and credentials (authorized) |
| Coercer / Responder | Forced authentication / poisoning |
## References
- `references/ad-attack-paths.md`
- `../pentest-tools/references/network-attack-defense.md`
- `../attack-chain/`
- seeds: `field-journal/seed-005_ad-certipy-esc1.md` `seed-007_ntlm-relay-coercer.md` `seed-013_kerberoasting-spn.md`
## Routing Context
**Upstream**: MASTER R24
**Downstream**: reporting `docs-generator`; EDR research needed → `edr-bypass-re`
**MUST NOT**: unauthorized DCSync / golden tickets against production
## Task Completion Self-Check
- [ ] Graph/enumeration before exploitation?
- [ ] Reproducible commands recorded and sanitized?
- [ ] Scope prohibitions respected?
- [ ] Checklist?
Files in this skill
- SKILL.md
- agents/openai.yaml
- references/ad-attack-paths.md
Attribution
Comments
Loading comments…
Newsletter
Get the weekly list of safe new skills, and the ones we pulled.