Detections
Explore public detection logic contributed by the community across SIEM and rule languages.
519 detections
Filters
Last updated
All Time
Detection languages
271
88
34
34
26
Contributors
144
82
54
25
22
Categories
347
256
244
69
63
Platforms
401
213
110
16
9
Products / Services
148
112
100
30
30
MITRE Techniques
18,123
15,639
12,581
8,394
6,215
CVEs
14
4
3
3
3
IDS Classtypes
7
2
1
1
IDS Protocols
11
Detects a Python process, potentially renamed, injecting code into the InstallUtil.exe utility. This behavior is a known TTP used by a Vietnamese threat actor in PureRAT campaigns, where a Python loader fetches a payload and injects it into a legitimate .NET utility to execute shellcode and evade defenses.
Detects the execution of a Python script with a specific command-line pattern used to inject AsyncRAT into memory. The pattern involves an input binary file, a key file argument, and a text file containing the key. This behavior is associated with a multi-stage AsyncRAT campaign.
Detects the execution of a Python script with a specific command-line pattern used to inject AsyncRAT into memory. The pattern involves an input binary file, a key file argument, and a text file containing the key. This behavior is associated with a multi-stage AsyncRAT campaign.
Detects a Python process executing a binary from an in-memory file descriptor via the /proc filesystem. This technique was observed in the malicious 'sympy-dev' PyPI package, which used it to execute a cryptominer from memory, bypassing some file-based detections.
Detects the execution of a Python interpreter (python.exe or pythonw.exe), potentially renamed, from a non-standard, user-writable location.
This can be an indicator of a threat actor dropping a portable Python environment to execute malicious scripts, as seen in campaigns like AsyncRAT.
This can be an indicator of a threat actor dropping a portable Python environment to execute malicious scripts, as seen in campaigns like AsyncRAT.
This rule detects patterns indicative of cryptomining malware execution, specifically targeting those distributed via impersonated PyPI packages. It looks for Python processes exhibiting characteristics such as: 1. Command lines containing keywords related to mining, cryptocurrency, hashing, and mining pools. 2. Python processes making suspicious network connections to stratum+tcp or mining pools. 3. Python processes engaging in resource-intensive operations like multiprocessing or threading, often utilizing all available CPU cores. 4. Python processes executing obfuscated code, potentially involving base64, zlib, or gzip decoding, combined with network requests. The rule triggers if any of these conditions are met, or if a combination of suspicious network activity and resource-intensive behavior, or obfuscated execution and resource-intensive behavior, is observed.
Detects the execution of a Python script from a non-standard user directory, using command-line parameters consistent with a tool used for Asynchronous Procedure Call (APC) injection. This specific command was observed in an AsyncRAT campaign where a script named 'ne.py' was used to inject a payload ('new.bin') into the explorer.exe process.
Detects AI/ML model configuration files (e.g., config.json, model_config.yaml) containing a suspicious '_target_' value, which could indicate an attempt to exploit a Hydra instantiation vulnerability for remote code execution (RCE).
Detects when a WebKit-based browser process on macOS spawns a shell or script interpreter. This is highly suspicious and can indicate successful exploitation of a browser vulnerability for initial code execution.
Detects when AppLocker prevents the execution of an Application, DLL, Script, MSI, or Packaged-App from running.
Detects python.exe executing from an unusual directory (e.g., a user's 'Contacts' folder) and injecting code into a benign process like notepad.exe. This is a specific behavior observed in the SERPENTINE#CLOUD campaign, where a Python-based loader is used to execute shellcode in memory.
Detects the execution of a Python interpreter with suspicious characteristics, such as running from a temporary location or using command-line arguments associated with obfuscation and in-memory execution. This behavior is consistent with Python-based stealers like Xillen Stealer.
Detects Executables in the Downloads folder without FileVersion,Description,Product,Company likely created with py2exe
Detects the LANDFALL Android spyware components, specifically the 'b.so' loader and the 'l.so' SELinux policy manipulator.
This rule detects the execution of Python web servers via the command line interface (CLI) on Linux systems. Adversaries often use Python's built-in HTTP server modules (like `http.server` or `SimpleHTTPServer`) in post-exploitation scenarios to quickly establish a web server without needing to install additional software. This allows them to easily transfer files between a compromised host and attacker-controlled systems, facilitating data exfiltration or further payload delivery.
This detection identifies network connections to external DNS servers over TCP (port 53) initiated by common scripting engines like PowerShell, Python, or WScript. While DNS over TCP is a valid protocol, it is less common for standard queries and is often used by malware for command and control (C2) or data exfiltration through DNS tunneling.
This rule detects HTTP requests containing User-Agent strings associated with the UNC6395 data theft campaign targeting Salesforce instances.
Detects execution of python using the "-c" flag. This could be used as a way to launch a reverse shell or execute live python code. The rule filters out common legitimate uses of `python -c` by Python itself, pip, and VS Code.
Detects the initial PowerShell download cradle associated with ClickFix/CORNFLAKE.V3 campaigns. This activity is often the result of a user being tricked into pasting and running a command from a malicious website.
Detects potential exploitation of FortiWeb CVE-2025-25257, a pre-authentication SQL injection vulnerability. This rule identifies an SQL injection attempt against the '/api/fabric/device/status' endpoint, followed by either a suspicious Python file drop in a library directory or a request to a specific CGI script used to trigger the payload.
Detects the use of the 'mdimport' command to index sensitive user directories. This is a key step in the Sploitlight TCC bypass technique (CVE-2025-31199), where a malicious Spotlight importer is triggered to run within the 'mdworker' process and exfiltrate data from TCC-protected locations by writing it to the system log.
Page 23 of 25