Detections
Explore public detection logic contributed by the community across SIEM and rule languages.
2,588 detections
Filters
Last updated
All Time
Detection languages
1,194
459
310
204
161
Contributors
487
455
323
171
123
Categories
1,884
937
415
400
360
Platforms
1,349
560
343
329
257
Products / Services
601
314
306
286
271
MITRE Techniques
18,119
15,639
12,589
8,399
6,227
2,588
CVEs
9
5
5
5
2
IDS Classtypes
142
96
49
11
6
IDS Protocols
177
49
44
17
4
Detects DNS resolutions for major AI/LLM provider domains (e.g., OpenAI, Anthropic, Google AI) originating from processes that are not common, known browsers or development tools. This behavior is indicative of potentially malicious applications or malware using AI APIs for automated tasks, such as content exfiltration or sophisticated C2 communication.
This rule detects potential data exfiltration activity where PowerShell is used to compress files followed by an outbound network connection to a known C2 IP address and port. It also monitors for direct outbound connections from PowerShell involving suspicious keywords indicative of data staging and exfiltration such as 'Compress-Archive', 'client.zip', and 'UPLOAD'.
This rule detects potential data exfiltration attempts by monitoring for network connections to URLs containing suspicious substrings (bit, onion, .cc, .su) over common web and DNS ports (53, 80, 443, 8080). It flags high-frequency unique connection attempts from a single device, which may indicate command-and-control (C2) traffic or data exfiltration over covert channels.
Correlates the reading of sensitive local files (id_rsa, .env, config.json) followed by an immediate POST request to an LLM API (specifically api.openai.com), indicating data theft via prompt injection or automated analysis.
Detects activity consistent with the Grandoreiro banking trojan, including DLL sideloading, screen capturing, keystroke hooking, unauthorized network connections to known banking infrastructure, and suspicious data exfiltration bursts.
Detects potential financial data exfiltration by monitoring for bulk downloads of files with financial keywords from SaaS/Office365 platforms, and identifies unauthorized processes attempting to access sensitive financial application data files (.qbw, .qbb, etc.) on local systems.
This rule monitors for three categories of potentially suspicious network behavior: connections to common CDN providers over non-standard ports, high-volume HTTPS connections directly to IP addresses (IP-direct) bypassing standard domain resolution, and high-volume traffic to CDN infrastructure that may indicate domain fronting or C2 communication. Such patterns are often used by adversaries to mask egress traffic or exfiltrate data.
Detects the execution of known command-line sync tools (Rclone, MEGAsync) or commands attempting to sync or move data to various cloud storage services (s3, gdrive, onedrive, mega, dropbox). The rule focuses on executions from suspicious or temporary directory paths and excludes processes signed by trusted publishers to minimize noise.
Detects programmatic file downloads from SharePoint and OneDrive using automation tools like PowerShell, PnP, Graph SDK, Python, Node.js, or generic CLI tools (curl, wget) rather than typical browser access. This can indicate data exfiltration.
This rule detects suspicious activity related to high-volume Server Message Block (SMB) share enumeration and the execution of a specific Python script named 'backup_dfs.py'. It combines three detection logics: identifying processes executing 'backup_dfs.py', detecting an unusual number of SMB network connections to common administrative shares (SYSVOL, NETLOGON, DFS), and flagging high-volume SMB-related identity directory events targeting SYSVOL or NETLOGON. This behavior could indicate an adversary performing discovery of network shares, collecting data, or preparing for data exfiltration.
This rule detects DNS queries with suspicious characteristics that may indicate DNS tunneling, data exfiltration, or command and control activity. It identifies queries with unusually long names, long first labels combined with high entropy, or a high number of subdomains combined with a long first label. Exclusions are made for common legitimate DNS suffixes and specific service-related domains.
This rule detects successful configuration modifications to diagnostic settings or storage services where the destination resource is identified as being outside of the organization's trusted subscriptions or expected storage naming conventions. This activity may indicate an adversary attempting to exfiltrate logs or data to an attacker-controlled storage account.
Detects network events where FTP (port 21) traffic contains specific timestamp-like patterns in the RemoteUrl, which could indicate an attempt to exploit the Squidbleed vulnerability related to timestamp-to-filename parsing anomalies. The rule looks for 'LIST' command combined with '12:34:56', '2026-', or 'timestamp' in the URL.
This rule detects the creation of a 'logs.dat' file, which is often associated with Remcos RAT for data exfiltration, in common user-related directories (AppData, Temp, Users, Documents) with a file size greater than or equal to 100KB. This pattern suggests an attempt to stage or exfiltrate collected data.
Detects browser connections to the Bing Search by Image endpoint, which can be exploited as a Server-Side Request Forgery (SSRF) proxy to bypass Content Security Policies (CSP) and exfiltrate data, as seen in the SearchLeak attack. The rule specifically looks for connections to 'bing.com/images/searchbyimage' that also include the 'imgurl=' or 'imgurl%3d' parameter, indicating an attempt to use Bing's server-side image fetching capability.
This rule detects network connections to known IP addresses associated with the FortiBleed vulnerability exploitation. These IP addresses have been categorized based on their observed roles in FortiBleed attacks, including C2, pentest lab, credential validation, and sniffer nodes. Monitoring these connections can help identify compromised systems communicating with FortiBleed infrastructure.
Detects outbound network connections to major Large Language Model (LLM) API endpoints (OpenAI, Google, Anthropic) initiated by processes not explicitly identified as standard web browsers or command-line utilities. This pattern may indicate automated data exfiltration ('prompt stealing' or unauthorized access to corporate data) using custom or malicious tools executing from suspicious directories like temp, appdata, or public folders.
This rule detects network connections to known IP addresses associated with the FortiBleed vulnerability exploitation. These IP addresses have been categorized based on their observed roles in FortiBleed attacks, including C2, pentest lab, credential validation, and sniffer nodes.
This rule detects a suspicious sequence of network connections indicative of information stealer activity, specifically targeting behaviors observed with malware like VVS Stealer. It identifies instances where a non-browser process, originating from common user-writable or temporary directories (AppData, Temp, Downloads, Users\Public), first connects to an external IP discovery service (ipify.org) and then subsequently communicates with Discord services (discord.com or discordapp.com) within a ten-minute window. This pattern suggests that malware is collecting the victim's public IP address before exfiltrating stolen data (e.g., account or browser information) to Discord infrastructure, often used as a command and control or data exfiltration channel by stealers.
Detects DNS lookups and network connections to known AI API endpoints (OpenAI, Anthropic, Google Generative Language) initiated by non-standard processes. This rule excludes common web browsers, productivity applications, and development tools to identify potentially unauthorized or suspicious interaction with AI services from internal processes.
Detects DNS lookups and network connections to known AI API endpoints (OpenAI, Anthropic, Google Generative Language) initiated by non-standard processes. This rule excludes common web browsers, productivity applications, and development tools to identify potentially unauthorized or suspicious interaction with AI services from internal processes.
Page 62 of 124