Detections
Explore public detection logic contributed by the community across SIEM and rule languages.
956 detections
Filters
Last updated
All Time
Detection languages
486
122
65
63
57
Contributors
328
210
131
82
29
Categories
504
436
202
184
173
Platforms
748
270
203
29
24
Products / Services
318
197
48
42
42
MITRE Techniques
18,123
15,639
12,580
8,394
6,215
956
CVEs
4
3
2
2
2
IDS Classtypes
6
6
4
3
IDS Protocols
15
4
Detects unauthorized processes (those not identified as standard web browsers) accessing browser profile directories (Local Storage or IndexedDB) associated with major AI platforms such as OpenAI, ChatGPT, Anthropic, and Copilot. This behavior is indicative of credential theft or session token exfiltration from local browser storage.
This rule detects the creation or modification of specific files associated with known malicious activity or remote access tools, specifically looking for shortcut files (LNK) named 'ApplianceConfig', 'ServerConfig', 'CredentialsConfig', or AnyDesk configuration files ('anydesk_svc.conf', 'anydesk_usr.conf'). These file patterns are often used by threat actors for persistence or to facilitate remote control.
This rule detects when processes other than known web browsers (Chrome, Firefox, Edge, Opera) attempt to access sensitive browser-related files such as 'Login Data', 'Cookies', and database files. It specifically flags instances where the accessing process or its parent process originates from 'C:\ProgramData' or where the parent process is 'powershell.exe', which is a common pattern for malicious credential access tools.
Detects a process that is not a standard office, browser, or system application accessing a high volume of document files (doc, docx, pdf, xlsx, xls, pptx, txt, odt) within user directories over a short time window. This behavior is indicative of potential automated file collection or data staging by unauthorized processes.
Detects the creation or modification of a large file named 'logs.dat' (100KB or greater) within common user-related directories such as AppData, Temp, Users, or Documents. This pattern can indicate data staging for exfiltration, where an adversary collects sensitive information into a single file before attempting to move it off the system.
Detects the unauthorized access of sensitive files (browser data, SSH keys, AWS credentials) followed by the creation of archives using common utilities (7zip, tar, zip) by processes spawned within npm/node.exe/powershell/wscript/chost.exe postinstall execution chains. This behavior is associated with credential theft and data staging for exfiltration, notably seen in campaigns like Operation Friday Harvest.
Detects access or creation of 'mcp.json' files within the '.amazonq' directory immediately following a 'git clone' operation. This sequence may indicate potential unauthorized access or exfiltration of Amazon Q assistant configurations or cached data associated with development environments.
Detects malicious behavior attributed to the Miasma/TonRAT malware, specifically targeting node.js or bun.exe runtime environments. The rule identifies suspicious network connections to command and control (C2) domains, access to sensitive developer credential files (e.g., .npmrc, .aws/credentials), execution of identified credential-harvesting functions in process command lines, and the execution of specific suspicious preinstall scripts.
Form-grabbing and credential stealing malware-as-a-service
Form-grabbing and credential stealing malware-as-a-service
Credential and browser data stealer distributed via OXLOADER campaigns
Detects suspicious access to wermgr.exe involving Alternate Data Streams (ADS) using the WDFOO tag, or access via Volume Shadow Copy (VSS) paths, often indicative of exploitation attempts related to CVE-2026-50656 where an adversary uses oplocks for race conditions against Windows Defender.
Detects anomalous process behavior characterized by rapid access to multiple sensitive credential files or enumeration of critical environment variables (e.g., cloud, AI, or cryptocurrency wallet keys). This behavior aligns with the mass credential harvesting techniques observed in the IronWorm/Shai-Hulud campaign.
Commodity .NET keylogger and remote access trojan
Detects node.exe copying browser history files to temporary directories, associated with the protocal.cjs backdoor.
Detects the execution of FTK Imager (FTKImager.exe), a forensic imaging tool that can be used by adversaries to perform data collection, including capturing physical memory, logical drives, or disk images from a local system.
Detects processes that access a threshold of five or more sensitive configuration, credential, or browser-stored login files within a single execution context, which is indicative of credential harvesting or automated collection activity.
Detects instances where a node or node.exe process accesses common shell history files (such as .bash_history, .zsh_history, etc.). This behavior is often associated with unauthorized access to command history, which may contain sensitive information like credentials or configuration details.
Detects instances where Node.js or NPM processes (or their parent processes) access sensitive files such as SSH private keys, AWS credentials, environment files, or configuration files containing secrets. This behavior is often associated with infostealers or malicious scripts exfiltrating credentials from developer environments.
Detects node.exe processes accessing sensitive AWS configuration files (credentials, config), .npmrc files, or suspicious data parsing utilities, indicative of the Polymarket Trap npm infostealer attempting to exfiltrate cloud credentials and tokens from developer workstations.
Detects node.exe (Node.js runtime) process initiating file read operations on sensitive files such as SSH keys, AWS credentials, Docker configuration files, GPG keyrings, or package manager configuration files (npmrc/pypirc). This behavior is characteristic of malicious npm packages or infostealers attempting to exfiltrate developer credentials, as observed in the Polymarket Trap supply chain campaign.
Page 30 of 46