A curated collection of Sigma rules to detect malicious behavior, adversary techniques, and indicators of compromise across different threat categories — including Remote Access Trojans (RATs), Advanced Persistent Threats (APTs), and Active Directory abuse.
Sigma is a generic rule format for log events that enables the sharing of detection logic across different SIEM and log analysis platforms. This repository provides Sigma-based threat detection rules that can be converted into queries for popular tools like:
- Splunk
- Elastic (Kibana)
- Sentinel
- Graylog
- QRadar
threat-detection/
├── APT/ # Sigma rules related to APT actor behavior and TTPs
├── Active Directory/ # Sigma rules for AD attacks like DCShadow, Kerberoasting, etc.
├── RAT/ # Sigma rules for detecting RATs such as Adwind, DarkMe, etc.
| Category | Example Rules |
|---|---|
| RAT | Adwind initial execution, DarkMe RAT execution |
| APT | Specific TTPs used by known APT groups |
| Active Directory | Privilege abuse, persistence, lateral movement tactics |
Each rule is mapped (when possible) to MITRE ATT&CK tactics and techniques.
-
Clone this repository:
git clone https://github.com/rodanmaharjan/threat-detection.git cd threat-detection -
Browse by category:
cd RAT/ -
Convert Sigma rules to your SIEM format using sigmac:
sigmac -t splunk -c config/splunk-windows.yml rule.yml
title: Adwind RAT Initial Execution
id: a1b2c3d4-5678-90ab-cdef-1234567890ab
description: Detects Java-based Adwind RAT execution via suspicious command-line usage.
status: experimental
author: rodanmaharjan
date: 2025/01/12
logsource:
product: windows
category: process_creation
detection:
selection:
CommandLine|contains:
- "javaw.exe"
- "-jar"
- "adwind"
condition: selection
fields:
- CommandLine
- ParentImage
- Image
level: high
tags:
- attack.execution
- attack.t1059
- rat.adwindContributions are welcome! If you'd like to submit a rule, ensure:
- It's in valid Sigma YAML format
- Includes appropriate metadata (e.g., MITRE mappings, tags)
- Tested or peer-reviewed before PR
This project is licensed under the MIT License.