Detections
Explore public detection logic contributed by the community across SIEM and rule languages.
11,831 detections
Filters
Last updated
All Time
Detection languages
5,623
1,822
1,261
860
752
Contributors
3,246
2,040
1,233
910
828
Categories
20,443
11,831
5,875
5,151
4,864
Platforms
9,731
2,539
1,169
276
100
Products / Services
3,075
2,976
2,352
1,065
955
MITRE Techniques
6,774
3,899
2,669
2,013
1,826
CVEs
21
19
18
14
14
IDS Classtypes
75
31
26
19
19
IDS Protocols
102
68
10
5
3
Detects anomalous process execution and loopback network activity associated with the Model Context Protocol (MCP) 'exec_in_session' method. This pattern indicates an adversary potentially abusing the AI coding assistant's MCP server capabilities to execute arbitrary commands as a C2 channel.
The following analytic identifies a Windows command interpreter process being executed where it's process name does not match it's original file name attribute.
Processes that have been renamed and executed may be an indicator that an adversary is attempting to evade defenses or execute malicious code.
Processes that have been renamed and executed may be an indicator that an adversary is attempting to evade defenses or execute malicious code.
Detects the ClickFix/fake-CAPTCHA initial access pattern where users are socially engineered into copying and pasting malicious commands into the Windows Run dialog or a browser-spawned shell. The rule matches on process creation events involving mshta.exe, powershell.exe, or wscript.exe initiated by explorer.exe or common shell-related processes, filtering out known administrative and deployment software.
The following analytic detects the execution of a DCRat "forkbomb" payload, which spawns multiple cmd.exe processes that launch notepad.exe instances in quick succession. This detection leverages Endpoint Detection and Response (EDR) data, focusing on the rapid creation of cmd.exe and notepad.exe processes within a 30-second window. This activity is significant as it indicates a potential DCRat infection, a known Remote Access Trojan (RAT) with destructive capabilities. If confirmed malicious, this behavior could lead to system instability, resource exhaustion, and potential disruption of services.
The following analytic detects potentially malicious processes that initiate a ping delay using an invalid IP address.
It leverages data from Endpoint Detection and Response (EDR) agents, focusing on command-line executions involving "ping 0 -n".
Malware like NJRAT was observed using this technique to introduce time delays for evasion tactics, such as delaying self-deletion.
If confirmed malicious, this activity could indicate an active infection attempting to evade detection, potentially leading to further compromise and persistence within the environment.
It leverages data from Endpoint Detection and Response (EDR) agents, focusing on command-line executions involving "ping 0 -n".
Malware like NJRAT was observed using this technique to introduce time delays for evasion tactics, such as delaying self-deletion.
If confirmed malicious, this activity could indicate an active infection attempting to evade detection, potentially leading to further compromise and persistence within the environment.
Detects the spawning of shells, interpreters, or system reconnaissance binaries by primary Citrix NetScaler ADC/Gateway processes (httpd, nsppe, aslearning). This behavior is strongly indicative of post-exploitation activity, such as remote command execution, often following the exploitation of vulnerabilities like CVE-2026-88771.
Detects Microsoft Office applications (Word, Excel, Outlook, PowerPoint) spawning common LOLBins or command-line interpreters. This behavior is frequently associated with malicious macros or exploitation attempts delivering secondary payloads.
Detects ClickFix-style attacks where a user is tricked into copying and pasting malicious, often obfuscated, commands into the Windows Run dialog or a command prompt. The detection identifies suspicious parent processes (like explorer.exe) spawning shell interpreters (powershell.exe, cmd.exe, mshta.exe) with inline encoded payloads or remote content retrieval patterns. It correlates these process executions with recent entries in the Windows RunMRU registry key to identify commands triggered via the Run dialog.
Detects instances where the Windchill Java/Tomcat application server process spawns command-line interfaces such as cmd.exe, powershell.exe, or bash.exe. This activity is indicative of post-exploitation behavior associated with CVE-2026-12569, where an attacker exploits a deserialization vulnerability to execute arbitrary commands, often for the purpose of dropping web shells.
Detects historical network connections, URL clicks, browser launches, email URLs, and process command lines associated with Wazza phishkit infrastructure IOCs including boegl-krysl.eu, coppietersrecherche.eu, and beacon-surge-sync.workers.dev.
Detects historical network connections, URL clicks, browser launches, email URLs, and process command lines associated with Wazza phishkit infrastructure IOCs including boegl-krysl.eu, coppietersrecherche.eu, and beacon-surge-sync.workers.dev.
Detects a behavioral chain where PowerShell stages a hidden base64-encoded command and schedules a VBScript task, or PowerShell fetches an upgrade.txt file via web client methods for execution.
Detects PowerShell execution performing port checks across multiple internal hosts using Test-NetConnection in a looping construct with output redirected to a log file, indicative of post-compromise reconnaissance (such as LONEPAGE scanning).
Detects PowerShell execution performing port checks across multiple internal hosts using Test-NetConnection in a looping construct with output redirected to a log file, indicative of post-compromise reconnaissance (such as LONEPAGE scanning).
Detects the execution of MATCHBOIL malware variants (PlannerAssistantManager.exe or RegularExpressionExplorer.exe) using specific command-line arguments (-auto, -plans, -renew) which trigger C&C logic.
Detects a potential web shell deployment by identifying a web server process creating or modifying a file in common web directory locations, followed shortly (within 10 minutes) by the execution of that same file.
Detects potentially malicious use of rundll32.exe by monitoring for execution from suspicious directory paths (e.g., Temp, AppData, Downloads), execution with arguments missing common legitimate entry points, unexpected network connections originating from rundll32.exe, or rundll32.exe spawning child processes (excluding known safe binaries like conhost.exe or werfault.exe).
Detects evidence of exploitation and post-exploitation activity related to CVE-2026-59310, an unauthenticated RCE vulnerability in the VMware vCenter Syslog Server. The rule monitors for suspicious child processes spawned by vmsyslogcollector, directory traversal patterns in command lines, and persistence mechanisms involving SSH authorized_keys modification, cron job creation, or systemd unit deployment on the VCSA shell.
Detects a sequence of events where a user clicks a shared link to an AI chatbot conversation followed closely (within 15 minutes) by the execution of potentially malicious processes (mshta, powershell, cmd) that include suspicious command-line patterns indicative of ClickFix-style social engineering.
This rule detects potential command injection attacks targeting the Zimbra 'swatchdog' service, specifically involving the abuse of SNMP trap notifications (CVE-2026-73570). The attack works by injecting malicious shell metacharacters into the 'zimbra-MIB::zmServiceName' argument, which is processed by a Perl-invoked 'swatchdog' script. The rule monitors process execution events for shell commands spawned by the swatchdog utility that contain suspicious shell operators and the targeted MIB parameter.
Detects awk-based shell command construction on ESX BusyBox that interleaves
an esxcli/vim-cmd command string with invisible Unicode code points (zero-width space,
zero-width non-joiner/joiner, word joiner, Mongolian vowel separator, left/right-to-left
marks, or Plane-16 private-use characters generated via sprintf hex escapes), strips them
with gsub(), and passes the result to system(). CrowdStrike documented this as one of 21
validated ESX obfuscation techniques that leaves only the invisible-character-laden string
in shell logs while the executed command is fully hidden from cleartext review.
an esxcli/vim-cmd command string with invisible Unicode code points (zero-width space,
zero-width non-joiner/joiner, word joiner, Mongolian vowel separator, left/right-to-left
marks, or Plane-16 private-use characters generated via sprintf hex escapes), strips them
with gsub(), and passes the result to system(). CrowdStrike documented this as one of 21
validated ESX obfuscation techniques that leaves only the invisible-character-laden string
in shell logs while the executed command is fully hidden from cleartext review.
Page 24 of 564