Guide · Agent skill
How to use the Windows Active Directory Security agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The windows-ad skill gives an AI agent a structured playbook for security testing of Windows Active Directory, the directory service that manages user accounts and access in many corporate networks. It is aimed at penetration testers, red team members and security researchers working under written authorization. After this guide you will know what the skill changes about the agent's behaviour, how to install it, how to phrase a request so it triggers, and what you need before using it.
Windows Active Directory Securityzhaoxuya520/reverse-skill · MIT Download the skill Quick start
- 1Confirm authorization. Get written permission and define the scope, including whether domain controllers and poisoning or relaying are allowed. The skill begins with this check.
- 2Download the ZIP. Download the windows-ad skill ZIP from this page.
- 3Install the skill. Unzip it into ~/.claude/skills/ for personal use or into a project's .claude/skills/ folder, or upload it under Settings → Capabilities → Skills in the Claude apps.
- 4Install the tools. Make sure Python and the tools the workflow uses, such as bloodhound-python, Impacket, NetExec, Certipy and BloodHound, are available in your environment.
- 5Start with enumeration. Ask the agent to map identities and BloodHound paths using the windows-ad skill before any exploitation is discussed.
- 6Record evidence and clean up. Keep reproducible, sanitized notes for each step and follow the cleanup step of the workflow.
What the windows-ad skill makes the agent do
An agent skill is a folder of instructions that an AI agent loads when a request matches the skill's description. This one covers Active Directory and Windows identity attack paths, including Kerberos, AD Certificate Services (AD CS), BloodHound attack paths, NTLM relay and domain privilege escalation research.
The skill does not make the agent act on its own. It sets a order of work: first confirm the authorization scope, then map identities and BloodHound attack paths, and only then consider exploitation. It asks the agent to record evidence at each step and to respect scope restrictions.
- Start with an authorization scope check, including whether domain controllers are in scope and whether poisoning or relaying is allowed.
- Enumerate identities and build BloodHound attack paths before attempting any exploitation.
- Work through topics such as Kerberoasting, AS-REP roasting, delegation abuse, AD CS template misconfigurations, ACL abuse and NTLM relay.
- Log evidence for each step and keep commands reproducible and sanitized.
- Lists common tools such as Impacket, NetExec, Certipy, BloodHound, Rubeus, Mimikatz, Coercer and Responder.
Install the windows-ad skill
Download the ZIP for the skill from this page. Unzip it into one of two locations. For a personal setup, use ~/.claude/skills/. For a single project, put it in that project's .claude/skills/ folder.
If you use the Claude apps instead, you can upload the skill under Settings → Capabilities → Skills. The skill's main file is SKILL.md, and it includes a reference file at references/ad-attack-paths.md.
- Personal: ~/.claude/skills/ (unzip the skill folder here).
- Project: .claude/skills/ inside the project folder.
- Claude apps: Settings → Capabilities → Skills, then upload.
How to trigger the skill with natural requests
You do not need a special command. The skill's description covers authorized Active Directory and Windows identity attacks, so ask for that kind of work in plain language and name the environment and your authorization. Stating the scope early matters, because the skill's first step is the scope check.
Good requests say what you are doing and what you are allowed to do. Vague requests with no scope are likely to get a scope-check question back, which is the intended behaviour.
- Using the windows-ad skill, help me plan an authorized assessment of the corp.local domain. Our written scope covers the domain controllers but not poisoning.
- Using the windows-ad skill, map BloodHound attack paths from this enumeration output before suggesting any exploitation.
- Using the windows-ad skill, review AD CS certificate templates for ESC1-style misconfigurations in our lab and write up the evidence.
Example workflow the skill follows
The skill's workflow begins with enumeration. Its example commands use NetExec over SMB and bloodhound-python to collect data, and both need domain credentials and authorization. The skill then lists common attack paths as a checklist, prioritising graph-based discovery over noisy action.
The reference file summarises each path with its prerequisite and the tool it points to. The reference also sets the core order: authorization, enumeration, path scoring, minimal verification, then cleanup.
- Kerberoast or AS-REP roast: service accounts with SPNs, or accounts that do not require pre-authentication; tools include GetUserSPNs, GetNPUsers and Rubeus.
- ESC1: a template that enrollees can register and that allows a forged SAN; tool is Certipy.
- ESC8: HTTP enrollment combined with relay; tool is ntlmrelayx.
- ACL to Domain Admin: GenericAll on a user or group; found with BloodHound.
- NTLM relay: applies where signing is not enforced; uses Responder and a relay.
nxc smb <range> -u user -p pass
bloodhound-python -d domain.local -u user -p pass -c All -ns <DC>What you need before using it
The skill assumes you already understand Kerberos, NTLM and Active Directory concepts and the tools it names. It also needs Python and the tools used by the workflow, valid domain credentials, and written authorization for the target environment.
The SKILL.md refers to other files in the wider repository, namely field-journal, attack-chain and pentest-tools. Those files are not included in the reviewed files, so the references to them will not resolve if you install only this skill. Some instructions are written in Chinese, so a reader who does not read Chinese should expect to check those parts.
- Python and tools such as bloodhound-python, Impacket, NetExec (nxc), Certipy, Rubeus, Mimikatz, Coercer or Responder.
- Valid domain credentials and written authorization for the target environment.
- Background in Kerberos, NTLM and Active Directory.
Safety notes: scripts and network access
This skill runs scripts and uses the network. It drives credential theft and privilege escalation tools, so it can cause real harm or illegal access if pointed at a domain you are not permitted to test. Do not use it on any domain without written permission.
Some listed actions, such as DCSync and golden tickets, can disrupt live systems. The skill itself says not to use unauthorized DCSync or golden tickets against production. Agree a scope that excludes production domain controllers unless you have explicitly agreed to test them, and check that the agent has stopped at each high-risk step for your confirmation.
- Confirm written authorization and scope before any command runs.
- Do not run DCSync or golden ticket actions against production without agreed scope.
- Keep evidence and sanitize credentials in your notes.
- Review every command the agent proposes before it executes on the network.
Tips for getting useful output
Give the agent enumeration output rather than asking it to guess the environment. The skill is designed to start from identity mapping and graph paths, so supplying BloodHound data or scan results makes its answers more specific.
Ask for a checklist and evidence log in the same request. The skill's self-check asks whether enumeration came before exploitation, whether commands were recorded and sanitized, and whether scope prohibitions were followed. Use those questions to review the agent's work.
- Provide BloodHound or enumeration output as input.
- Ask for the self-check checklist to be filled in.
- Name the specific attack family (for example Kerberoasting or AD CS) to keep the answer focused.
What people use it for
- Planning an authorized penetration test of an Active Directory domain with a scoped checklist.
- Reviewing BloodHound attack paths from collected data and documenting each step.
- Checking AD CS certificate templates for misconfigurations in a lab environment.
- Studying Kerberoasting and AS-REP roasting weaknesses in a controlled research lab.
- Structuring a red team write-up that records evidence for each identity attack path.
Frequently asked questions
How do I install the windows-ad agent skill?
Download the ZIP from this page and unzip it into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder. You can also upload it in the Claude apps under Settings → Capabilities → Skills.
Is the windows-ad skill free to use?
The skill is listed under the MIT licence. The source provided does not mention pricing, so check the licence text in the repository for the terms that apply.
What do I need to run the windows-ad skill?
You need Python and tools such as bloodhound-python, Impacket, NetExec, Certipy, Rubeus, Mimikatz, Coercer or Responder, plus valid domain credentials and written authorization. The skill also assumes you understand Kerberos, NTLM and Active Directory.
Can I use this skill on any Active Directory domain?
No. It is for authorized testing only. Do not use it against a domain you do not have written permission to assess, and avoid production domain controllers unless the scope explicitly covers them.
Why does the skill not run my attack immediately?
The skill is built to start with a scope check, then enumeration and BloodHound paths, and to avoid destructive exploitation first. This is intended behaviour, so state your scope clearly in your request.
Why are some referenced files missing?
The SKILL.md points to field-journal, attack-chain and pentest-tools files in the wider repository. Those files are not included in the reviewed files, so install the rest of the repository if you need them.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use the windows-ad skill if you are a penetration tester, red team member or security researcher with written authorization for an Active Directory engagement and want a structured, graph-first checklist. Pick something else if you have no authorization or security background, or if you need a tool that works without the Kerberos, NTLM and AD knowledge the skill assumes.
Download the skill