Detections
Explore public detection logic contributed by the community across SIEM and rule languages.
3,324 detections
Filters
Last updated
All Time
Detection languages
1,225
935
481
239
184
Contributors
626
512
469
337
261
Categories
20,443
11,831
5,875
5,151
4,864
3,324
Platforms
1,411
1,220
873
142
138
Products / Services
973
444
393
392
360
MITRE Techniques
1,525
1,085
885
593
540
CVEs
42
28
26
26
24
IDS Classtypes
390
220
127
40
26
IDS Protocols
671
150
30
18
12
This rule looks for URL patterns matching common SQLi keywords used to break out of the array key for CVE-2026-9082.
This rule detects a successful Remote Desktop Protocol (RDP) login that is immediately preceded by 20 or more failed RDP authentication attempts from the same source IP to the same destination IP within a 15-minute window. This pattern is indicative of a brute force attack that has successfully compromised credentials.
Detects suspicious modifications to Notepad++ configuration files (config.xml or shortcuts.xml) by processes other than Notepad++ itself. This behavior may indicate an attempt to exploit vulnerabilities (e.g., CVE-2026-48778 or CVE-2026-48800) by injecting unvalidated interpreter commands, leading to arbitrary code execution or other malicious activities.
This rule detects the loading of known vulnerable drivers (BYOVD) or drivers from suspicious user-writable paths, combined with the termination of security products. This combination of events is highly indicative of an adversary attempting to gain kernel-level privileges and disable defenses.
Detects an NGINX worker process spawning a shell or interpreter process, the post-exploitation primitive for CVE-2026-42945 (NGINX Rift). The exploit chain corrupts the ngx_pool_t cleanup handler and triggers system() with attacker-controlled shell commands, causing the nginx worker to fork /bin/sh with the -c flag. Since system() always invokes /bin/sh -c, the highest-fidelity match is nginx → /bin/sh -c <command>. Generalizes to catch arbitrary interpreter spawning (bash, python, perl, ruby, netcat variants, socat) which would result from any future post-exploitation pivot. Scoped to Linux because the exploit relies on the fork-based multi-worker architecture (Linux/BSD only). Windows nginx is not affected by the heap feng shui exploitation primitive.
Identifies potentially malicious client-side redirects or forced OAuth flows initiating from the Outlook Web Access (OWA) interface, characterized by 302 statuses and redirect parameters. This activity may indicate post-exploitation actions related to CVE-2026-42897.
Detects unusual JavaScript-related URI query parameters in Exchange IIS logs accessing the OWA directory. This may indicate attempted exploitation of the CVE-2026-42897 XSS vulnerability.
Detects a temporal pattern characteristic of CVE-2026-42897 XSS exploitation: an OWA user reads a message (via Bind, OpenMessage, or similar EWS/OWA endpoints), and within 60 seconds, the same session issues a POST request to create or modify an inbox rule. This rapid succession from the same session cookie indicates programmatic action (XSS-driven JavaScript) rather than human-driven workflow. The rule correlates events by client IP and username within a 60-second window.
Detects workloads created with privileged mode, host namespace sharing, dangerous Linux capabilities, Docker socket, containerd socket, or sensitive hostPath mounts. These configurations can be abused by adversaries to escape containers and gain access to the underlying host.
Detects the execution of 'explorer.exe' from outside the standard C:\Windows\ directory when launched by Notepad++, indicating potential exploitation of the CVE-2026-25926 unsafe search path vulnerability.
This rule detects unusual network connection patterns characterized by a high volume of connections to a remote IP address within a 5-minute window, combined with anomalous connection durations. Specifically, it flags scenarios where there are more than 1000 connections and either the 95th percentile of connection durations is very short (less than 50ms) or the 5th percentile of connection durations is very long (greater than 30000ms). These patterns can indicate activities such as network scanning, denial-of-service attempts, or command and control communication.
Detects the execution of Notepad++ using the '-settingsDir=' argument, which can be used to redirect the application to load attacker-controlled configuration files, potentially exploiting vulnerabilities like CVE-2026-48778. This technique could allow an adversary to achieve persistence or execute malicious code by manipulating Notepad++'s configuration.
Detects the exploitation primitive of CVE-2026-3102 at the post-injection execution layer. The exploit succeeds when exiftool constructs a /usr/bin/setfile command line containing attacker-injected shell metacharacters (single quotes for command injection, backticks or $() for command substitution, semicolons for command chaining). Detecting setfile invocations spawned by exiftool with these characters in arguments is the smoking-gun signal. Even future variants of this vulnerability class (or related n-day exploits in the same family) would produce this same parent/child + metacharacter pattern.
Docker Run/Exec Guardrails Bypass
Cortex XDR
This rule detects the execution of a Docker command with the '--privileged' flag on a Linux system, combined with the execution of specific system information gathering commands. This combination is often indicative of an attempt to escape a container or gain elevated privileges within a containerized environment, potentially leading to host compromise. The '--privileged' flag grants a container nearly all capabilities of the host machine, making it a critical security concern when used maliciously.
Detects `git rebase` commands containing the `--exec=` flag when spawned by a Gogs or Gitea server process. This activity is indicative of the exploitation of an argument injection vulnerability (CVE-2022-46153) in Gogs/Gitea, which allows an authenticated user to achieve Remote Code Execution (RCE) by creating a pull request with a specially crafted branch name. The rule specifically looks for `git` or `git.exe` processes executing `rebase` with `--exec=` and originating from `gogs`, `gitea`, or common shell processes that might be spawned by them.
Monitors Cisco Catalyst SD-WAN Controller logs for successful administrative logins. This may indicate exploitation of an authentication bypass vulnerability granting unauthorized admin access.
Detects Cisco SD-WAN control connection state changes indicating a new 'vmanage' peer connection from an unauthorized IP address. This rule is designed to identify potential exploitation of vulnerabilities like CVE-2026-20182, where attackers establish unauthorized infrastructure ties by creating new vManage peer connections. The rule filters for syslog messages from the Viptela-vSmart-VDAEMON indicating a 'control-connection-state-change' with 'new-state:up' and 'peer-type:vmanage', and then excludes connections from a list of authorized vManage IP ranges.
Detects the execution of the public Proof-of-Concept for the "Dirty Frag" Linux Kernel Privilege Escalation vulnerability. It monitors for the specific command-line flags '--force-esp' and '--force-rxrpc', as well as the execution of /bin/sh or /bin/bash with the TERM=xterm environment variable, which is indicative of the PoC's payload.
This rule detects potential exploitation of CVE-2026-4892 by correlating inbound DHCPv6 traffic (UDP/547) with a dnsmasq process crash (SIGSEGV/SIGABRT exit codes 132, 134, 135, 139) on the same host within a 60-second window. This vulnerability involves sending crafted DHCPv6 messages with oversized Client Identifier options to a dnsmasq server configured with --dhcp-script, leading to a denial of service.
Detects the registration of the PLURIBUS Cloud Files sync provider used by the CVE-2023-36036 (cldflt.sys heap overflow) privilege escalation exploit in the Rapid7-documented ModeloRAT incident. The exploit registers a fake sync provider with the specific GUID {904EE598-0511-4664-82A8-22C4A7501044} pointing to %TEMP%\cldflt\, then writes a placeholder file (Link.log) to trigger the vulnerable code path. The provider GUID, name, and registry path are all unique to this exploit variant — legitimate sync providers (OneDrive, Dropbox, Box) use vendor-specific GUIDs and never register under %TEMP%.
Detects dnsmasq or its DHCP helper child process terminating abnormally on Linux hosts due to signals like SIGSEGV, SIGABRT, SIGBUS, or SIGILL. This behavior is indicative of a crash, potentially caused by exploitation attempts of vulnerabilities such as CVE-2026-4892, a heap buffer overflow in dnsmasq's helper.c when processing oversized DHCPv6 Client Identifier (CLID) options. This rule fires post-impact, complementing network-side detections.
Page 111 of 159