Emiliano Mema
@NosalvaGreeceTrusted contributorCompletionist
0 followers634 downloads589 copies6 likes1,947 views
282 detections
Filters
Last updated
All Time
Detection languages
159
62
43
18
Categories
95
52
49
34
29
Platforms
156
52
34
25
17
Products / Services
74
55
40
40
19
MITRE Techniques
81
76
34
31
22
CVEs
2
2
1
IDS Classtypes
22
14
4
3
IDS Protocols
22
9
5
5
1
Detects the abuse of Visual Studio Code (and compatible IDEs like Cursor) workspace trust to trigger malicious tasks upon opening a folder. This pattern is associated with the 'Void Dokkaebi' (Famous Chollima) threat group's propagation of the 'DEV#POPPER' RAT. The rule identifies instances where the code editor spawns common scripting or network utilities (like PowerShell, curl, or Node.js) with command-line arguments indicative of file downloads or execution, as defined in a local '.vscode/tasks.json' file.
Detects developer tooling processes such as Node.js, npm, and yarn—or processes spawned by code editors like VS Code and Cursor—establishing network connections to specific blockchain API endpoints (Tron, Binance Smart Chain, Aptos). This behavior is associated with the DEV#POPPER RAT, which utilizes these blockchain transaction channels to stage encrypted payloads and retrieve configuration keys.
Detects the use of osascript to generate fake macOS system-style dialog boxes that solicit user credentials or administrator privileges. This technique is often employed by macOS infostealers to conduct GUI-based credential harvesting or to bypass user interaction requirements for privilege escalation.
Detects the creation of a new domain user followed by its addition to sensitive, high-privileged groups (e.g., Domain Admins, Enterprise Admins) using the native Windows net.exe or net1.exe utilities. This pattern aligns with known post-compromise activity observed in campaigns like Bumblebee, AdaptixC2, and Akira ransomware, where attackers create stealthy, elevated accounts (e.g., masquerading as backup service accounts) to maintain persistence and full domain control.
Detects malicious markdown files (SKILL.md) that impersonate agent instruction files. These files contain hidden-content constructs (like CSS display:none, HTML comments, or invisible Unicode characters) combined with common command execution patterns (like curl, wget, or PowerShell encoded commands), characteristic of infostealer campaigns targeting AI agent platforms.
Detects the execution of the 'chmod +x' command to mark a file named '/tmp/helper' as executable, followed by the execution of that same binary within a 5-minute window. This behavior is indicative of an attacker dropping and executing a payload from a temporary directory.
Detects the execution of PowerShell with encoded commands, a common technique used by attackers to obfuscate malicious scripts and evade detection.
Detects reconnaissance patterns associated with SynkLoader, specifically focusing on in-memory PowerShell execution used to harvest system information, user privileges, running services, process listings, and Active Directory computer counts.
Detects outbound HTTP POST requests characterized by an 'application/octet-stream' content type, a specific URI structure consisting of three components (20-40 characters, alpha-numeric, 20-40 characters), and an encrypted payload. This behavioral pattern is associated with SynkLoader malware command-and-control (C2) communications.
Detects the execution of Google Chrome or Microsoft Edge in headless mode with the Chrome DevTools Protocol (CDP) remote debugging port enabled. This configuration allows external control of the browser instance and is often abused by attackers to perform browser-based attacks, intercept traffic, or exfiltrate session data.
Page 5 of 29