Ad Opsec Telemetry logo

Ad Opsec Telemetry

Community
ADScanPro
ad-opsec-telemetry

The telemetry each Active Directory technique generates and what alerts a defender: Kerberoasting produces Event 4769 with RC4 encryption (0x17) and an MDI alert, DCSync produces Event 4662 with the DS-Replication-Get-Changes GUID, AS-REP roasting produces Event 4768 with no pre-auth, LSASS dumping is blocked by EDR, plus a lateral-movement telemetry table by protocol (SMB/WMI/WinRM/RDP/DCOM). Use this whenever you run or plan an offensive AD technique and need to know what noise it makes, when writing the engagement's detection notes, or when a defender wants to know what to monitor. Teaches red teamers what is loud and blue teamers what to watch. Standard-tooling knowledge, no vendor engine.

Overview

PublisherADScanPro
RepositoryClaude-AD
Skill namead-opsec-telemetry
Stars
215
Forks
34
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by ADScanPro on GitHub. Read the source before you install it.

Installation

Install the Ad Opsec Telemetry AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/ADScanPro/Claude-AD/tree/main/skills/ad-opsec-telemetry
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/ADScanPro/Claude-AD.git /tmp/Claude-AD
mkdir -p .claude/skills
cp -r /tmp/Claude-AD/skills/ad-opsec-telemetry .claude/skills/ad-opsec-telemetry
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Ad Opsec Telemetry in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Ad Opsec Telemetry on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Ad Opsec Telemetry is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

AD OPSEC and Telemetry

Every technique leaves a trace. The goal here is not evasion. It is knowing the noise profile of each action so you can document it for the client before the engagement, pick the quieter of two functionally equivalent techniques, and correlate what you did with what their SOC saw. For a defender, read the same tables backwards: they are the events to monitor and alert on.

Two rules of engagement throughout:

  • Document the noise. Tell the client what each high-signal action generated, with the event ID and the source IP, so they can find it in their logs.
  • Coordinate the loud ones. DCSync, mass spraying, coercion at scale, and LSASS dumps need explicit client sign-off or an end-of-engagement window when detection no longer matters.

Microsoft Defender for Identity (MDI) is the sensor that matters most in AD; it reads DC traffic directly and ships tuned detections for most of what follows.


Kerberoasting: Event 4769, RC4 (0x17)

Requesting service tickets for accounts with SPNs, to crack offline.

  • Event 4769 (Kerberos service ticket requested) fires for each SPN targeted. The tell is Ticket Encryption Type = 0x17 (RC4-HMAC). Attackers request RC4 because the resulting hash cracks fastest, but a service that normally uses AES suddenly requested with RC4 is the classic signature.
  • MDI: "Suspected Kerberoasting attack" is on by default and triggers on the volume and RC4 pattern.
  • Requesting all SPNs at once amplifies the signal. Targeting one high-value account is far quieter than roasting the whole domain.
bash
GetUserSPNs.py corp.local/user:pass -dc-ip <dc_ip> -request -outputfile roast.txt
  • MITRE: T1558.003 (Steal or Forge Kerberos Tickets: Kerberoasting).
  • Defender watch: 4769 with 0x17 for accounts that otherwise use AES; a single principal requesting many distinct SPN tickets in a short window.
  • Remediation: strong (25+ char) service-account passwords or gMSA; AES-only on service accounts; monitor 4769 by encryption type.

AS-REP Roasting: Event 4768, no pre-auth

Cracking accounts that have Kerberos pre-authentication disabled, no valid credential needed to request the roastable material.

  • Event 4768 (TGT requested) with Pre-Authentication Type = 0. Normal accounts always pre-authenticate; a 4768 with no pre-auth is the signature.
  • MDI: "Suspected AS-REP Roasting attack."
bash
GetNPUsers.py corp.local/ -usersfile users.txt -dc-ip <dc_ip> -no-pass -format hashcat
  • MITRE: T1558.004 (AS-REP Roasting).
  • Defender watch: 4768 with pre-auth type 0; any account carrying DONT_REQUIRE_PREAUTH.
  • Remediation: remove DONT_REQUIRE_PREAUTH wherever possible; strong passwords on accounts that genuinely need it; alert on the flag being set.

DCSync: Event 4662, replication GUIDs

Replicating credentials out of the DC using directory-replication rights. High severity, never transparent, so coordinate before running.

  • Event 4662 (operation on a directory object) with the replication rights GUIDs:
    • 1131f6aa-9c07-11d1-f79f-00c04fc2dcd2: DS-Replication-Get-Changes
    • 1131f6ad-9c07-11d1-f79f-00c04fc2dcd2: DS-Replication-Get-Changes-All
  • MDI: "Suspected DCSync attack (replication of directory services)": a critical alert in any competent SOC.
  • The source host in the DC logs is your machine's IP, not a real DC. That mismatch (a non-DC requesting replication) is itself the detection.
bash
secretsdump.py corp.local/user:pass@dc01.corp.local -just-dc-user krbtgt
  • MITRE: T1003.006 (OS Credential Dumping: DCSync).
  • Defender watch: 4662 with the Get-Changes GUIDs from any principal that is not a DC; replication requests sourced from non-DC IPs.
  • Remediation: audit and minimize who holds Get-Changes / Get-Changes-All; alert on replication from non-DC sources; tier-0 isolation.

LSASS dumping: blocked by EDR

Dumping process memory of lsass.exe to extract credentials.

  • Modern EDR in blocking mode (CrowdStrike Falcon, SentinelOne, Defender for Endpoint) intercepts MiniDumpWriteDump against lsass. The result is a blocked/empty dump, a crashed process, or an immediate alert naming your tool.
  • Event 4688 (process creation) plus EDR-specific handle-to-lsass telemetry.

Alternatives that avoid touching lsass live:

  • SAM/SYSTEM/SECURITY registry hive dump (local admin) → offline parsing.
  • DPAPI offline extraction.
  • Volume Shadow Copy of the hive + offline parse.
  • gMSA / LAPS fetch over LDAP (no endpoint touch at all).
bash
# Registry-hive route: avoids a live lsass handle
secretsdump.py -sam sam.save -system system.save LOCAL
  • MITRE: T1003.001 (LSASS Memory).
  • Defender watch: process handles opened to lsass with read/VM-read access; EDR memory-access alerts.
  • Remediation: Credential Guard; LSASS as a protected process (RunAsPPL); EDR in blocking mode.

BloodHound collection: a recognizable LDAP signature

  • Collector queries have well-known shapes: (objectCategory=computer) pulling dozens of attributes, plus queries for msDS-AllowedToDelegateTo, delegation and ACL attributes in bulk.

  • MDI and legacy ATA carry specific detection for BloodHound-style enumeration.

  • Quieter collection requests only the attributes it needs (never *) and spreads queries over time instead of firing them all at connect.

  • MITRE: T1087 (Account Discovery), T1069 (Permission Groups Discovery).

  • Defender watch: a single principal issuing large attribute-heavy LDAP sweeps in a short window.


Lateral movement: telemetry by protocol

When you need remote execution, the protocol you pick determines the noise. From loudest to quietest:

ProtocolTelemetryOPSEC note
PsExec (SCM service install)Event 7045 service install — very visible, routinely EDR-blockedAvoid when EDR is active
WMI execEvent 4688 process create — MDE detectsQuieter than PsExec
Scheduled taskEvent 4698/4702 task created — moderately visibleAcceptable
WinRM / PSRemotingLegitimate channel, but Event 4624 logon type 3 + PowerShell logsPreferred when available
DCOM / MMCFewer known signaturesBest profile under EDR
  • MITRE: T1021 (Remote Services) and subtechniques; T1569.002 (Service Execution); T1053.005 (Scheduled Task); T1047 (WMI).
  • Rule: record which protocol you used and its profile in the report, so the client can correlate with their endpoint logs.

Coercion and NTLM relay: high-volume auth noise

  • Coercion (PetitPotam, PrinterBug, DFSCoerce) forces a target to authenticate to you, producing inbound auth requests visible in SIEM. Coercing at scale generates hundreds of auth events.

  • ntlmrelayx in automatic mode can capture credentials of real users in production, so coordinate before running it against a live environment.

  • MITRE: T1187 (Forced Authentication), T1557.001 (LLMNR/NBT-NS Poisoning and SMB Relay).

  • Defender watch: spikes of inbound authentications to a non-standard host; EFSRPC / spoolss / DFS RPC calls to unexpected destinations.

  • Remediation: SMB signing enforced; LDAP channel binding; EPA on AD CS web endpoints; disable spooler on DCs.


SMB null sessions: reconnaissance signal

  • Windows Server 2016+ disables null sessions by default; some SOCs alert on the attempt as anonymous recon.

  • When you hold credentials, always use them. Null session is a last-resort fallback only, and worth documenting when used.

  • MITRE: T1135 (Network Share Discovery), T1087.


Actions that require explicit coordination

Show a warning and get sign-off (or document that it ran) before any of these:

ActionWhy
DCSyncCritical MDI alert; your IP visible in DC logs
Password sprayingRisk of locking out real accounts
ntlmrelayx (auto)May capture real users' credentials in production
Coercion at scaleHundreds of auth requests, SIEM-visible
Kerberoasting all accountsMDI volume alert; document targeted accounts
LSASS dumpEDR-detectable; risk of crashing the process

How to use this

Running a technique: check its row, warn the client about the events it will generate, and prefer the quieter equivalent when there is one (WMI over PsExec, one SPN over the whole domain, registry-hive over live lsass). Writing the report: turn these tables into the detection-and-remediation section so the defender can find every action you took and close the gap that let it work.

Frequently asked questions

What does the Ad Opsec Telemetry AI skill do?

The telemetry each Active Directory technique generates and what alerts a defender: Kerberoasting produces Event 4769 with RC4 encryption (0x17) and an MDI alert, DCSync produces Event 4662 with the DS-Replication-Get-Changes GUID, AS-REP roasting produces Event 4768 with no pre-auth, LSASS dumping is blocked by EDR, plus a lateral-movement telemetry table by protocol (SMB/WMI/WinRM/RDP/DCOM). Use this whenever you run or plan an offensive AD technique and need to know what noise it makes, when writing the engagement's detection notes, or when a defender wants to know what to monitor. Teach...

Why use Ad Opsec Telemetry on TypingMind?

Because you install it once and use it with any model. Ad Opsec Telemetry is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Ad Opsec Telemetry in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/ADScanPro/Claude-AD/tree/main/skills/ad-opsec-telemetry. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Ad Opsec Telemetry?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Ad Opsec Telemetry?

As many as you like. As long as a model supports skills, you can use Ad Opsec Telemetry with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Ad Opsec Telemetry AI skill free?

Yes. It is published on GitHub by ADScanPro under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇