Detections
Explore public detection logic contributed by the community across SIEM and rule languages.
4,534 detections
Filters
Last updated
All Time
Detection languages
2,471
532
343
301
286
Contributors
1,303
815
737
215
193
Categories
20,439
11,829
5,873
5,151
4,864
4,534
Platforms
3,833
1,048
471
39
29
Products / Services
1,829
1,138
211
183
171
MITRE Techniques
1,598
1,416
1,297
744
473
CVEs
19
17
16
11
11
IDS Classtypes
23
1
IDS Protocols
16
7
1
Detects unauthorized modifications to the .ssh/authorized_keys file. Such changes are often performed by adversaries to establish persistence by ensuring their own public keys allow them remote access to the host. The rule filters out known configuration management and system tools to reduce noise.
Detects the deployment of the Hermes Agent AI framework by the Carbonato botnet. The rule identifies the creation or modification of a 'SOUL.md' persona file within specific directories (/root/.hermes/ or /opt/netd/) followed by an outbound network connection to a known LLM gateway IP or the Telegram API. This pattern is indicative of a post-exploitation AI agent being installed for command and control or task execution.
Detects a behavioral pattern associated with the NeedyMantis first-stage loader, which involves the loading of specific suspicious DLLs (masquerading as common software components) followed by the extraction or creation of an 'encryptbase64.ps1' second-stage loader within a short timeframe (10 minutes) on the same host.
Detects unauthorized file creation within the /root/.hermes/loot/ directory and specific path combinations associated with the CARBONATO botnet's Hermes Agent. This activity is indicative of credential harvesting, specifically the exfiltration of AI API keys for various LLM providers.
Detects potential persistence mechanisms on Windows systems by monitoring for the creation or modification of Registry 'Run' or 'RunOnce' keys, as well as the creation of files within the Windows Startup folder. These actions are commonly used by adversaries to ensure malicious code executes automatically upon user login or system startup.
Detects a suspicious pattern where an AI coding agent or user creates a public GitHub repository with specific naming conventions (e.g., 'sweeper-demo', 'gitshot-images') and subsequently pushes content containing sensitive terms like 'screenshot' within a short timeframe. This behavior is indicative of potential data exfiltration of internal development screenshots to a public repository.
Detects a suspicious pattern where an AI coding agent or user creates a public GitHub repository with specific naming conventions (e.g., 'sweeper-demo', 'gitshot-images') and subsequently pushes content containing sensitive terms like 'screenshot' within a short timeframe. This behavior is indicative of potential data exfiltration of internal development screenshots to a public repository.
Detects the creation, modification, or renaming of common JSP web shell filenames (x.jsp, u.jsp, u2.jsp) within known web application directories (e.g., webapps, Peoplesoft, WebLogic). This pattern is indicative of an adversary attempting to establish a web shell for persistent access or command execution on a vulnerable web server.
Detects Gamaredon (Armageddon) group activity characterized by Microsoft Office applications (winword.exe or excel.exe) writing to global template files (Normal.dotm or Personal.xlsb) to implant malicious macros, or dropping .vbs or .lnk files into startup or template directories to achieve persistence and facilitate further propagation.
Detects the installation of known AI-powered coding assistant extensions (e.g., Copilot, Tabnine, Cursor) on a device, followed by significant source-code repository cloning or archival activity by Git within two hours. This pattern is potentially indicative of intellectual property theft or sensitive source code exfiltration facilitated by a newly introduced AI tool.
Detects the execution of PowerShell with a bypass execution policy that performs a download or web request to save content to a file in the user's temp directory, followed by the immediate execution of that file. This pattern is characteristic of multi-stage malware droppers or fileless attack techniques attempting to stage and execute malicious scripts from temporary locations.
Detects evidence of potential adversarial interference with Windows Defender updates and MRT.exe (Malicious Software Removal Tool). The rule correlates high volumes of temporary file creation in user directories with unauthorized access to MRT.exe by third-party processes and repeated Windows Defender update failures, indicating a potential attempt to disrupt endpoint security operations.
Detects the creation or writing of executable binary files (e.g., .exe, .dll, .bat, .ps1, .vbs, .scr) to sensitive Windows administrative shares (ADMIN$, C$, IPC$) via SMB. This behavior is indicative of lateral movement where an attacker attempts to stage tools or scripts on a remote system for subsequent execution.
Detects creation or modification of files under /etc/sudoers or /etc/sudoers.d/ by a
process outside sanctioned package- and configuration-management tooling. Writing a
passwordless sudo rule is the finishing move in CIFSwitch (CVE-2026-46243) and OVSwrap
(CVE-2026-64531) and a common root-persistence step generally. Enriched with writer
context (non-root, writable-path, shell one-liner, NOPASSWD) and a heuristic RiskScore /
TriagePriority, plus DeviceInfo host context (MachineGroup, OS, logged-on users).
Detection logic adapted from Elastic Security Labs - Linux LPE detection framework.
process outside sanctioned package- and configuration-management tooling. Writing a
passwordless sudo rule is the finishing move in CIFSwitch (CVE-2026-46243) and OVSwrap
(CVE-2026-64531) and a common root-persistence step generally. Enriched with writer
context (non-root, writable-path, shell one-liner, NOPASSWD) and a heuristic RiskScore /
TriagePriority, plus DeviceInfo host context (MachineGroup, OS, logged-on users).
Detection logic adapted from Elastic Security Labs - Linux LPE detection framework.
Detects a behavioral fingerprint associated with the NeedyMantis group, characterized by the creation of a DLL and a identically-named, extensionless, encrypted archive file in the same directory within a short time window. This approach identifies the underlying packaging strategy rather than relying on static file names or known paths.
Detects .ps1-named files dropped to disk that are never actually executed by a PowerShell interpreter (powershell.exe / pwsh.exe) within a short window afterward. NeedyMantis's second-stage loader (e.g. encryptbase64.ps1) is raw x64 shellcode, not a real script -- it's read and executed directly by the dropping process, so no genuine PowerShell host process ever references the file by name. Replaces an earlier version of this rule that watched DeviceImageLoadEvents for a .ps1 extension: Windows can only emit an image-load event for a file with a valid PE header loaded via the OS loader, and raw shellcode has no PE header and is never mapped that way, so that approach would not have fired on this malware. Caveat: legitimate deployment tooling that stages a script for delayed or remote execution can also produce this create/execute mismatch; tune the window and add known-good deployment-tool exclusions for your environment.
Detects a process obtaining handle access to the LSASS process (Sysmon Event ID 10) followed by the creation of a file (Sysmon Event ID 11) using an extension other than '.dmp'. This behavior is characteristic of adversaries attempting to obfuscate credential dumping activities by bypassing simple extension-based detection rules.
Detects the creation of files commonly associated with ransomware notes (e.g., README.txt, HOW_TO_DECRYPT.txt) across multiple directories, which is a strong indicator of the impact phase where files have been encrypted and the adversary is providing recovery instructions.
This rule correlates web clicks on Google Sites URLs containing keywords related to GlobalProtect with the subsequent creation or renaming of an unsigned or improperly signed GlobalProtect.msi file on the same endpoint within a 30-minute window. This behavior is indicative of a spearphishing attempt using a masqueraded landing page to deliver malicious or unauthorized software.
This rule detects the creation or presence of specific file artifacts ('PDF_C2089_20260911100446.exe' and 'active_desktop_render_x64.dll') within the '\AppData\Microsoft\Update\' directory. These file names and paths are characteristic of potential malware staging or persistence mechanisms, specifically associated with the SilverFox threat activity.
Detects malicious activity associated with the ClickFix/Vidar campaign, where adversaries create persistence using the Windows Run registry key while masquerading as the legitimate 'VoidTools Everything.exe' utility. The rule correlates registry modifications with the creation or execution of a file path mimicking the legitimate software in suspicious locations.
Page 12 of 216