Detections
Explore public detection logic contributed by the community across SIEM and rule languages.
6,537 detections
Filters
Last updated
All Time
Detection languages
3,970
998
598
428
233
Contributors
1,363
1,250
1,066
954
384
Categories
2,933
2,196
2,017
1,857
1,411
Platforms
39,936
7,266
6,537
3,866
3,647
Products / Services
4,107
2,418
1,587
270
254
MITRE Techniques
1,450
1,310
1,022
934
696
CVEs
46
46
33
18
12
IDS Classtypes
1,753
458
404
325
220
IDS Protocols
2,202
773
381
300
93
This rule detects a payload download after a redirect chain, which is a common technique used in user execution attacks. It specifically looks for HTTP responses containing a redirect pattern in the HTTP referer and then checks for the 'MZ' magic bytes (indicating a Windows executable) in the file data within a short depth. This suggests a user being redirected and then automatically downloading a malicious executable.
This rule detects the download of files that have a 'double extension' pattern, specifically where a common document or image file extension (pdf, doc, jpg, png) is followed by an executable extension (exe, bat, vbs, ps1). This pattern is often used by attackers to trick users into executing malicious files, as the operating system might only display the first extension, making the file appear benign. The detection is based on analyzing the HTTP URI for this specific naming convention.
This rule detects suspicious executable downloads from email-related referers. It specifically looks for HTTP traffic where the referer header contains keywords like 'mail.', 'webmail.', 'outlook.', 'gmail.', or 'yahoo.com', indicating the download originated from an email client or webmail interface. The rule further narrows down the detection to files containing the 'MZ' header (characteristic of Windows executables) with a size greater than 10000 bytes, suggesting a potential malicious executable being delivered via email.
This rule detects HTTP traffic containing the string "duckyScript" in the file data, indicating a potential payload delivery page associated with OMG Cables or Rubber Ducky devices. These devices are often used for injecting malicious keystrokes or other payloads.
This rule detects potential cloud API access using a stolen token by monitoring HTTP requests to AWS S3, Azure management, or Google APIs that contain an 'Authorization: Bearer' header. A threshold of 50 such requests within 10 seconds from the same source IP triggers the alert, indicating a possible brute-force or high-volume access attempt with a compromised token.
This rule detects HTTP POST requests to common cloud console login pages (AWS, Microsoft Azure, Google Cloud) that originate from a suspicious IP range. The rule uses a threshold of 5 attempts within 30 seconds from the same source IP to trigger an alert, indicating potential brute-force or credential stuffing attacks against cloud accounts.
This rule detects VNC (Virtual Network Computing) remote access sessions initiated from any source IP address to any IP address within the local network ($HOME_NET) on port 5900. It specifically looks for the 'RFB 00' string within the first 7 bytes of the TCP payload, which is characteristic of the VNC protocol handshake. A threshold is applied to trigger an alert if 3 such connections are observed from the same source IP within a 60-second window, indicating potential repeated or sustained VNC activity.
This rule detects suspicious software installers being downloaded over HTTP. It looks for HTTP responses containing the 'MZ' header (indicating a Windows executable) with a file size greater than 1MB, and where the HTTP URI contains keywords like 'setup', 'install', or 'update'. The rule is thresholded to trigger if three such events occur from the same source within 60 seconds, aiming to identify potential malware or unwanted software distribution.
This rule detects the download of potentially trojanized software updates. It specifically looks for HTTP traffic flowing to the client, where the URI contains "update" (case-insensitive), the HTTP host contains "update.", "cdn.", or "download." (case-insensitive), and the file data starts with "MZ" (indicating a Windows executable) within the first 2 bytes. Additionally, the downloaded file size must be greater than 500,000 bytes.
This rule detects attempts to download Python packages from the official Python Package Index (PyPI) where the package name in the URI appears to be a typosquatted version of common legitimate packages. Specifically, it looks for HTTP requests to pypi.org with URIs containing 'requets' (instead of 'requests'), 'boto' (instead of 'boto3' or similar), 'djang0' (instead of 'django'), 'flask2' (instead of 'flask'), or 'numpy2' (instead of 'numpy'). This indicates a potential supply chain attack where an attacker has registered a similarly named malicious package to trick users into downloading it.
This rule detects a potential watering hole attack where a legitimate website has been injected with a script that attempts to load content from an IP address. Specifically, it looks for HTTP responses (flow:to_client) containing '<script src=' followed by a URL where the source is an IP address (e.g., 'src="http://1.2.3.4/script.js"'). This technique is often used by attackers to compromise visitors of a trusted site by redirecting them to malicious content or exploiting vulnerabilities.
This rule detects potential drive-by exploit kit landing pages by inspecting HTTP response data for the presence of both "exploit" and "shellcode" (case-insensitive). It specifically looks for these keywords in the data flowing from the server to the client over an established HTTP connection. The rule is categorized as an exploit-kit detection and is associated with MITRE ATT&CK technique T1189 (Drive-by Compromise) and tactic Initial Access.
This rule detects potential rogue captive portal credential harvesting pages by inspecting HTTP responses for the presence of keywords like 'captive', 'password', and 'connect' within the file data. These keywords, when found together in a client-bound HTTP response, suggest an attempt to mimic a legitimate captive portal to steal user credentials.
This rule detects attempts to capture credentials via an 'Evil Twin' access point by looking for HTTP POST requests to URIs containing both 'login' and 'wifi'. This combination suggests a malicious captive portal designed to trick users into submitting their credentials.
Detects DNS requests to a known BNB Smart Chain Testnet JSON-RPC endpoint used by AnimateClipper to dynamically resolve its actual C2 domain via a smart contract.
Detects web requests containing common SQL injection payloads or specific PoC metacharacters embedded within JSON:API filter array keys targeting Drupal CVE-2026-9082. The rule specifically looks for requests to Drupal's JSON:API that include 'filter[' and '[condition][value][' in the URI query, combined with SQL injection indicators such as '--', 'UNION', 'SELECT', '$$', '.$', or 'RAPTORMARKER42'.
Detects network traffic over TCP port 886, which is specifically used by Atlas RAT for command and control check-ins, plugin downloads, and encrypted data exfiltration in TA4922 campaigns.
This rule detects network connections originating from clients using a specific User-Agent string associated with the SessionGate malware. The User-Agent string 'NSIS_InetLoad' combined with 'Mozilla' is indicative of SessionGate C2 check-ins, leveraging the NSIS (Nullsoft Scriptable Install System) InetLoad plugin. The rule searches across common proxy and web logs for this pattern.
Detects DNS requests to a known BNB Smart Chain Testnet JSON-RPC endpoint used by AnimateClipper to dynamically resolve its actual C2 domain via a smart contract.
Detects network connections to domains associated with Argamal RAT command and control infrastructure.
This rule detects potential DNS hijacking attempts via a compromised resolver. It specifically looks for UDP traffic on port 53 (DNS) to the home network, containing the DNS response code '81 80' (indicating a standard query response) and the PCRE pattern '/\x00\x01\x00\x00/' which corresponds to a single question in the DNS query. A threshold of 20 occurrences within 10 seconds from the same source IP triggers the alert, suggesting a high volume of suspicious DNS responses.
Page 247 of 312