Detections
Explore public detection logic contributed by the community across SIEM and rule languages.
295 detections
Filters
Last updated
All Time
Detection languages
116
41
34
33
20
Contributors
106
38
28
23
20
Categories
165
157
111
46
39
Platforms
229
188
103
9
9
Products / Services
10,660
9,755
6,689
4,418
3,748
295
MITRE Techniques
172
112
60
51
43
CVEs
6
6
3
2
2
IDS Classtypes
4
IDS Protocols
4
Detects common runtime processes (Node.js, npx, Python) often used for Model Context Protocol (MCP) servers accessing sensitive local credential files such as AWS credentials, SSH private keys, or environment files. This behavior is indicative of potential credential harvesting by unauthorized MCP servers in developer environments.
Detects the execution of the 'sckit' Go-based implant when spawned by Node.js or Python processes using the command-line arguments 'stage0 --config64'. This pattern is associated with malicious npm and PyPI supply chain packages (memtensor) designed to harvest credentials and establish persistence.
Detects the execution of the 'sckit' Go implant, a malicious binary associated with trojanized '@memtensor/memos-cloud-openclaw-plugin' or 'MemoryOS' packages. The rule identifies instances where this binary is spawned as a child process of a language runtime (Node.js or Python) when the execution originates from specific directory paths associated with these packages, or when the parent process command line indicates these packages are being initialized.
Detects execution of PowerShell or Python scripts characterized by common patterns found in AI-generated code, including specific obfuscation techniques (base64, string manipulation), boilerplate error handling structures, verbose comments, and generic variable naming conventions. High scores indicate a higher probability of automated or obfuscated script execution.
Detects instances where machine learning runtime processes (e.g., Python, torch-model-loader) load serialized model files (e.g., .pkl, .joblib, .pth) and subsequently spawn suspicious child processes like command shells, scripting engines, or download utilities. This behavior is indicative of arbitrary code execution via unsafe deserialization of a malicious model artifact.
Detects instances where Node.js, Python, or Go processes spawn command shell interpreters (cmd.exe, powershell.exe, etc.). This behavior is often indicative of C2 command execution by malware, such as the Graphalgo RAT, which may use these languages to execute shell-level commands after initial infection or payload execution.
Detects the execution of the 'sckit' Go implant, a malicious binary associated with trojanized '@memtensor/memos-cloud-openclaw-plugin' or 'MemoryOS' packages. The rule identifies instances where this binary is spawned as a child process of a language runtime (Node.js or Python) when the execution originates from specific directory paths associated with these packages, or when the parent process command line indicates these packages are being initialized.
Detects malicious prompt injection payloads targeting AI agent frameworks, specifically designed to bypass Semantic Kernel's filter blocklist. The payload utilizes Python reflection mechanisms by crawling the type hierarchy (starting from a tuple) to locate and invoke 'BuiltinImporter.load_module()', effectively bypassing filters that look for direct references to forbidden modules or functions like '__import__', 'eval', or 'os'. This allows for indirect arbitrary command execution within the agent's application context.
Detects the execution of known Python scripts (CES_Enroll.py, ntlm_ces_relay.py, ces_negotiate_ntlm.py) often associated with NTLM relay attacks or forced authentication techniques. These scripts leverage Python to interact with authentication protocols or relay requests.
Detects the execution of known Python scripts (CES_Enroll.py, ntlm_ces_relay.py, ces_negotiate_ntlm.py) often associated with NTLM relay attacks or forced authentication techniques. These scripts leverage Python to interact with authentication protocols or relay requests.
This rule detects potential remote code execution (RCE) attempts targeting AI agent applications, specifically those utilizing the Semantic Kernel framework. It identifies suspicious command-line patterns indicative of a Python type hierarchy traversal attack (CVE-2026-26030), which is used to bypass security blocklists within an eval() sink to execute arbitrary commands like 'os.system()'. The rule monitors for the spawning of common shell or utility processes by Python or .NET processes associated with AI agent activity.
Detects the execution of the 'sckit' Go implant, a malicious binary associated with trojanized '@memtensor/memos-cloud-openclaw-plugin' or 'MemoryOS' packages. The rule identifies instances where this binary is spawned as a child process of a language runtime (Node.js or Python) when the execution originates from specific directory paths associated with these packages, or when the parent process command line indicates these packages are being initialized.
Detects the execution of the 'sckit' Go implant, a malicious binary associated with trojanized '@memtensor/memos-cloud-openclaw-plugin' or 'MemoryOS' packages. The rule identifies instances where this binary is spawned as a child process of a language runtime (Node.js or Python) when the execution originates from specific directory paths associated with these packages, or when the parent process command line indicates these packages are being initialized.
Detects exploit attempts targeting a Python prompt-injection vulnerability (CVE-2026-26030). The rule identifies command lines attempting to traverse Python's object hierarchy (e.g., using __class__, __base__, __subclasses__) to import the os module and execute system commands. It also flags subsequent child processes (e.g., calc.exe, cmd.exe) spawned by Python interpreters or Semantic Kernel host processes.
Detects attempts to bypass input sanitization in AI agent frameworks, specifically Semantic Kernel, by using Python bracket or subscript notation to access sensitive attributes or functions (e.g., __class__, __subclasses__, __import__, eval, exec). This technique is characteristic of filter injection attacks aiming to achieve remote code execution in AI agent environments.
Detects potential exploitation of CVE-2025-3248 in Langflow by identifying suspicious inbound network traffic targeting the /api/v1/validate/code endpoint, correlated with subsequent python child processes containing base64 decoding and execution primitives such as eval, exec, or subprocess.
Detects bulk file creation of XML files associated with Dahua credential exports, likely generated by a Python script, which is indicative of credential harvesting or staging for exfiltration.
Detects the execution of the 'sckit' Go implant, a malicious binary associated with trojanized '@memtensor/memos-cloud-openclaw-plugin' or 'MemoryOS' packages. The rule identifies instances where this binary is spawned as a child process of a language runtime (Node.js or Python) when the execution originates from specific directory paths associated with these packages, or when the parent process command line indicates these packages are being initialized.
Detects Python processes (python.exe, python3.exe, pythonw.exe) performing DNS resolution for Microsoft identity, device login, or Graph API endpoints. This activity is consistent with automated tooling, such as python-requests, performing device-code flow polling or API token authentication within an environment.
Detects unauthenticated attempts to exploit CVE-2025-3248 in the Langflow 'validate/code' endpoint, characterized by missing authentication headers and the presence of suspicious Base64-encoded payloads or command execution keywords. Additionally, it identifies potential post-exploitation activity where Langflow processes spawn Python subprocesses.
Detects the scripted, non-human-paced registration of a rogue external authentication method by monitoring a sequence of specific Azure Audit Log operations (Add application, Add service principal, Add delegated permission grant) occurring within 60 seconds of each other by the same actor. The detection is further corroborated by analyzing Microsoft Graph Activity Logs for associated Python requests to specific Graph API URIs.
Page 2 of 15