Detections
Explore public detection logic contributed by the community across SIEM and rule languages.
11,847 detections
Filters
Last updated
All Time
Detection languages
5,637
1,822
1,261
860
754
Contributors
3,246
2,040
1,233
910
828
Categories
20,460
11,847
5,879
5,158
4,866
Platforms
9,745
2,545
1,173
276
100
Products / Services
3,075
2,985
2,353
1,065
955
MITRE Techniques
6,780
3,901
2,671
2,015
1,829
CVEs
21
19
18
14
14
IDS Classtypes
75
31
26
19
19
IDS Protocols
102
68
10
5
3
Detects a chained sequence of suspicious activity initiated by agentic AI development tools (e.g., Claude Code, Aider, AutoGen). The rule identifies the execution of these tools followed by local reconnaissance commands, lateral movement attempts, and outbound network connections originating from the same host, indicating potential automated exploitation or credential abuse.
Detects anomalous, chained execution patterns initiated by AI agents integrated with Model Context Protocol (MCP) servers. The rule identifies a multi-stage sequence involving the invocation of browser automation tools followed by shell execution and subsequent outbound network activity, potentially indicating an AI-driven attack chaining filesystem, shell, or cloud-API access in an unauthorized manner.
Detects coordinated activity patterns across multiple distinct tenant environments within a one-hour window. The rule looks for shared infrastructure (C2 domains), malicious tool hashes, or specific agentic-CLI command line signatures (e.g., related to orchestration, reconnaissance, or exfiltration) that suggest an AI-orchestrated actor or automated attack campaign affecting multiple organizations simultaneously.
Detects anomalous child process spawning from common web browsers or productivity applications (e.g., Office, Acrobat) on hosts that have recently communicated with known LLM or code-generation APIs. This behavioral correlation is intended to identify the potential execution of AI-generated exploit code.
Detects attempts to disable or tamper with Microsoft Windows Defender via registry modifications or malicious PowerShell/reg.exe commands, specifically targeting security settings like DisableAntiSpyware, DisableRealtimeMonitoring, and exclusion paths.
Detects a sequence of events where a potential multimedia lure (PDF, image, or media file) is created or downloaded via a web browser or communication client, followed by an AI agent (e.g., ChatGPT, Claude) accessing the file, and subsequently initiating a shell execution process (PowerShell, CMD, etc.) involving suspicious command-line patterns or references to the original file. This behavior is indicative of an AI-assisted prompt injection attack where malicious content is processed and executed through agent tooling.
Detects a sequence of events where a potential multimedia lure (PDF, image, or media file) is created or downloaded via a web browser or communication client, followed by an AI agent (e.g., ChatGPT, Claude) accessing the file, and subsequently initiating a shell execution process (PowerShell, CMD, etc.) involving suspicious command-line patterns or references to the original file. This behavior is indicative of an AI-assisted prompt injection attack where malicious content is processed and executed through agent tooling.
Detects a sequence of events where a potential multimedia lure (PDF, image, or media file) is created or downloaded via a web browser or communication client, followed by an AI agent (e.g., ChatGPT, Claude) accessing the file, and subsequently initiating a shell execution process (PowerShell, CMD, etc.) involving suspicious command-line patterns or references to the original file. This behavior is indicative of an AI-assisted prompt injection attack where malicious content is processed and executed through agent tooling.
Detects anomalous activity where AI assistant processes (e.g., Claude, ChatGPT, GitHub Copilot) execute a high volume of diverse discovery commands. The rule correlates multiple discovery categories, such as account, network service, system information, and network configuration discovery, occurring within a short timeframe to identify potential abuse of AI-integrated development tools for host reconnaissance.
Detects anomalous activity where AI assistant processes (e.g., Claude, ChatGPT, GitHub Copilot) execute a high volume of diverse discovery commands. The rule correlates multiple discovery categories, such as account, network service, system information, and network configuration discovery, occurring within a short timeframe to identify potential abuse of AI-integrated development tools for host reconnaissance.
Detects anomalous activity where AI assistant processes (e.g., Claude, ChatGPT, GitHub Copilot) execute a high volume of diverse discovery commands. The rule correlates multiple discovery categories, such as account, network service, system information, and network configuration discovery, occurring within a short timeframe to identify potential abuse of AI-integrated development tools for host reconnaissance.
Detects anomalous, rapid enumeration of system, network, process, and software configuration settings by AI coding assistant processes (e.g., Claude, Cursor, ChatGPT). This behavior often signifies an AI agent being coerced or configured to perform environment discovery or credential gathering, indicating potential compromise or abuse of the LLM-integrated development environment.
This rule detects potential AI prompt injection attacks delivered via email. It identifies users clicking external links directed at AI platforms (like OpenAI, Claude, or Perplexity) that contain suspicious URL parameters or instructions designed to manipulate or override the AI's standard behavior. It correlates the web request with a subsequent process execution on the host that includes suspicious keywords in the command line, suggesting an attempt to automate or exploit the AI assistant through local system interaction.
This rule detects potential AI prompt injection attacks delivered via email. It identifies users clicking external links directed at AI platforms (like OpenAI, Claude, or Perplexity) that contain suspicious URL parameters or instructions designed to manipulate or override the AI's standard behavior. It correlates the web request with a subsequent process execution on the host that includes suspicious keywords in the command line, suggesting an attempt to automate or exploit the AI assistant through local system interaction.
This rule detects potential AI prompt injection attacks delivered via email. It identifies users clicking external links directed at AI platforms (like OpenAI, Claude, or Perplexity) that contain suspicious URL parameters or instructions designed to manipulate or override the AI's standard behavior. It correlates the web request with a subsequent process execution on the host that includes suspicious keywords in the command line, suggesting an attempt to automate or exploit the AI assistant through local system interaction.
Detects anomalous discovery behavior originating from known AI agent and development tools (e.g., Claude, Cursor, Aider). The rule monitors for a rapid sequence of commands spanning multiple categories of enumeration (Network, File/Directory, Software, and System) within a short timeframe, which indicates an attempt to map the runtime environment, permissions, or system capabilities.
Detects anomalous discovery behavior originating from known AI agent and development tools (e.g., Claude, Cursor, Aider). The rule monitors for a rapid sequence of commands spanning multiple categories of enumeration (Network, File/Directory, Software, and System) within a short timeframe, which indicates an attempt to map the runtime environment, permissions, or system capabilities.
Detects anomalous discovery behavior originating from known AI agent and development tools (e.g., Claude, Cursor, Aider). The rule monitors for a rapid sequence of commands spanning multiple categories of enumeration (Network, File/Directory, Software, and System) within a short timeframe, which indicates an attempt to map the runtime environment, permissions, or system capabilities.
The following analytic detects potential exploitation of the CrushFTP authentication bypass vulnerability (CVE-2025-31161). This detection identifies suspicious command execution patterns associated with exploitation of this vulnerability, such as executing mesch.exe with specific arguments like b64exec or fullinstall. This activity is indicative of an attacker exploiting CVE-2025-31161 to gain unauthorized access to the CrushFTP server and perform post-exploitation activities.
Detects AI agent tooling (e.g., Copilot, Cursor, Ollama) that makes outbound network calls to known AI model providers followed by the execution of suspicious shell commands within a short duration. The rule flags high-severity activities such as download-cradles, credential reconnaissance, LOLBin usage, and persistence mechanisms. Informational alerts are generated for standard CLI activity within the same timeframe.
Detects instances where msiexec.exe, the Windows Installer utility, spawns cmd.exe to execute a batch file named 'xupdate.bat' located within the 'AppData\Roaming\Traiolx Custom Utils\' directory. This behavior is indicative of malicious persistence or execution mechanisms, often associated with loaders like LegionLoader that leverage installer binaries for evasion.
Page 59 of 565