Detections
Explore public detection logic contributed by the community across SIEM and rule languages.
2,294 detections
Filters
Last updated
All Time
Detection languages
1,284
250
199
142
138
Contributors
660
544
375
96
95
Categories
20,456
11,839
5,876
5,155
4,865
2,294
Platforms
1,505
929
381
70
51
Products / Services
743
407
225
164
140
MITRE Techniques
856
683
604
437
302
CVEs
12
10
9
9
8
IDS Classtypes
24
4
3
1
1
IDS Protocols
20
6
6
1
Detects anomalous multi-agent chaining where one AI-enabled agent instance initiates a connection to an AI API and subsequently hands off information to a second, distinct agent instance, which then performs an suspicious downstream action such as spawning a shell or initiating an unauthorized external network connection. This behavior is intended to identify potential abuse of AI-coding assistant tools for automated execution chains.
Detects anomalous multi-agent chaining where one AI-enabled agent instance initiates a connection to an AI API and subsequently hands off information to a second, distinct agent instance, which then performs an suspicious downstream action such as spawning a shell or initiating an unauthorized external network connection. This behavior is intended to identify potential abuse of AI-coding assistant tools for automated execution chains.
Detects suspicious force push events to the 'main' branch in GitHub repositories. The rule flags pushes that include hardcoded commit hashes associated with known malicious activity, pushes where the author and committer identities do not match, or pushes containing specific indicators such as 'indexe.cjs' or 'preinstall' scripts which are common in software supply chain attacks.
This rule detects potentially malicious modifications to package.json files in npm projects or suspicious npm registry telemetry. It flags manifest files containing common obfuscation markers or excessive/unusual dependencies often associated with supply chain attacks, such as crypto, child_process, and common third-party packages frequently used in malicious npm activity.
This rule detects modifications to Windows Registry run keys and startup folder locations that are commonly used for persistence. It specifically flags when entries point to scripts, interpreted binaries, or files located in directories often used by attackers to hide malicious payloads (e.g., AppData, Temp, ProgramData).
Detects unauthorized modifications to SSH authorized_keys files. Attackers often modify these files to inject their own public keys, ensuring persistent access to a compromised Linux host. This rule excludes known configuration management and provisioning tools to reduce noise.
KQL Query
Detects the modification of registry keys under HKCU\Software\Classes\CLSID\...\InProcServer32, which is a common method for achieving persistence. The rule flags when these registry keys point to file paths within the local user profile (AppData), which is frequently used by malicious actors to load custom DLLs when the corresponding COM object is initialized.
This rule detects potentially malicious activity involving Node.js modules. It monitors for the deletion of specific 'sharedLoad.min.js' or 'extended' files within 'node_modules' directories, the execution of Node.js processes attempting file deletion operations (like 'unlink' or 'rm') within 'node_modules', and the creation or modification of JavaScript files within 'node_modules' that contain 'btree' references. This activity may indicate an attempt to tamper with application dependencies, inject malicious code, or remove artifacts.
Detects a multi-stage, automated attack chain performed by AI-based development agents (e.g., Claude, Copilot, Cursor). The rule correlates process discovery, credential access, file collection (staging), and exfiltration over web services within a tight 15-minute window, identifying potential autonomous exploitation and data exfiltration. Covered T1046, T1552.001, T1078, T1005, T1567
Detects a multi-stage, automated attack chain performed by AI-based development agents (e.g., Claude, Copilot, Cursor). The rule correlates process discovery, credential access, file collection (staging), and exfiltration over web services within a tight 15-minute window, identifying potential autonomous exploitation and data exfiltration. Covered T1046, T1552.001, T1078, T1005, T1567
Detects a multi-stage, automated attack chain performed by AI-based development agents (e.g., Claude, Copilot, Cursor). The rule correlates process discovery, credential access, file collection (staging), and exfiltration over web services within a tight 15-minute window, identifying potential autonomous exploitation and data exfiltration. Covered T1046, T1552.001, T1078, T1005, T1567
Detects a multi-stage, automated attack chain performed by AI-based development agents (e.g., Claude, Copilot, Cursor). The rule correlates process discovery, credential access, file collection (staging), and exfiltration over web services within a tight 15-minute window, identifying potential autonomous exploitation and data exfiltration. Covered T1046, T1552.001, T1078, T1005, T1567
Detects a file named dnsapi.dll created/modified outside System32/SysWOW64/WinSxS (the only legitimate locations for the real system DLL). Now requires the file to BOTH land in one of the known NeedyMantis sideload staging folders AND match the published 3448-byte spoofed-config size -- previously the size check alone could match unrelated dnsapi.dll copies anywhere on disk; requiring both signals together removes that bypass.
Detects a file named dnsapi.dll created/modified outside System32/SysWOW64/WinSxS (the only legitimate locations for the real system DLL). Now requires the file to BOTH land in one of the known NeedyMantis sideload staging folders AND match the published 3448-byte spoofed-config size -- previously the size check alone could match unrelated dnsapi.dll copies anywhere on disk; requiring both signals together removes that bypass.
Detects a file named dnsapi.dll created/modified outside System32/SysWOW64/WinSxS (the only legitimate locations for the real system DLL). Now requires the file to BOTH land in one of the known NeedyMantis sideload staging folders AND match the published 3448-byte spoofed-config size -- previously the size check alone could match unrelated dnsapi.dll copies anywhere on disk; requiring both signals together removes that bypass.
Detects Impacket-style hands-on-keyboard staging: a NeedyMantis bundle file dropped by cmd.exe/wmiprvse.exe/services.exe carrying the distinctive remote-execution fingerprint (ADMIN$ share reference, %COMSPEC% invocation, or 2>&1 output redirection used by wmiexec/smbexec/atexec), followed within 15 minutes by execution referencing the same bundle folder with the same fingerprint. Requiring the ADMIN$/%COMSPEC%/redirection fingerprint on both the copy and the execution -- rather than just matching on common process names like cmd.exe or svchost.exe -- removes the bulk of ordinary software installation and update activity that also uses cmd.exe to stage files.
Detects behavior indicative of the MLTBackdoor malware, specifically the creation or modification of a DLL file (such as *dlp.dll) shortly after or before the creation of an RC4-encrypted 'data.bin' file on the same host. The rule also looks for an optional preceding archive extraction event of common file types from Temp or Downloads directories by standard extraction utilities, which may indicate the staging of the initial payload.
Detects the modification of AI agent configuration files (e.g., mcp.json, config.json) followed shortly by the agent spawning a child process that matches typical Model Context Protocol (MCP) server execution patterns. This behavior may indicate an attacker has modified the configuration to inject malicious server commands or tools that the AI agent will execute upon next initialization, effectively achieving persistent code execution via the agent's legitimate functionality.
Detects the modification of AI agent configuration files (e.g., mcp.json, config.json) followed shortly by the agent spawning a child process that matches typical Model Context Protocol (MCP) server execution patterns. This behavior may indicate an attacker has modified the configuration to inject malicious server commands or tools that the AI agent will execute upon next initialization, effectively achieving persistent code execution via the agent's legitimate functionality.
Detects the modification of AI agent configuration files (e.g., mcp.json, config.json) followed shortly by the agent spawning a child process that matches typical Model Context Protocol (MCP) server execution patterns. This behavior may indicate an attacker has modified the configuration to inject malicious server commands or tools that the AI agent will execute upon next initialization, effectively achieving persistent code execution via the agent's legitimate functionality.
Page 15 of 110