Detections
Explore public detection logic contributed by the community across SIEM and rule languages.
4,230 detections
Filters
Last updated
All Time
Detection languages
2,175
1,018
354
260
212
Contributors
1,013
818
688
574
212
Categories
20,456
11,839
5,877
5,156
4,865
4,230
Platforms
2,933
889
585
248
198
Products / Services
2,239
1,453
841
416
344
MITRE Techniques
1,164
1,087
870
798
439
CVEs
22
20
20
18
17
IDS Classtypes
927
408
222
146
104
IDS Protocols
1,914
105
70
47
6
Detects potential ML model-stealing attacks against inference API endpoints. The rule identifies anomalous behavior characterized by an extremely high volume of diverse queries from a single API key over a 24-hour period. Such patterns are consistent with systematic decision-boundary probing or input-space coverage attempts used for training a replica model (distillation or knockoff attacks) without direct access to model weights.
Detects systematic and high-frequency probing of internet-exposed LLM chatbot or API gateway endpoints. The rule monitors for sequences of requests, use of known prompt-infiltration or jailbreak testing toolkits, and varied path access indicative of automated reconnaissance or adversarial prompt engineering attempts.
Detects external reconnaissance activity targeting AI backend services. Adversaries scan for known AI infrastructure ports (such as TensorFlow Serving, Triton, or Ray) and common API paths (e.g., /v1/models, /predict) to identify potential targets. The detection utilizes proxy logs to correlate these suspicious network attributes, high request rates, or known scanning user agents.
Detects network activity associated with the AMOS (Atomic macOS Stealer) malware, specifically identifying outbound HTTP POST requests to the '/api/t' URI path occurring on non-standard ports. The detection utilizes a threshold-based trigger to identify potential C2 communication heartbeats.
This rule detects network activity involving specific domain names (e.g., velvet-otter-glagceis.life) or HTTP request patterns associated with malicious JavaScript or trackers, which may indicate C2 communication or malicious script loading.
This rule detects network traffic patterns indicative of the Lumma Stealer 'ClickFix' phishing campaign. It identifies outbound requests to suspicious domains ending in .cyou utilizing specific user agents (WinHttp or Microsoft BITS), which are characteristic of malicious PowerShell scripts attempting to retrieve secondary payloads.
This rule detects HTTP POST requests to .php files containing suspicious PHP execution functions in the request body, such as 'system', 'eval', or 'shell_exec'. These patterns are indicative of an attempt to execute commands via an in-memory or injected PHP web shell on an F5 BIG-IP APM device.
Detects post-exploitation activity against Citrix NetScaler appliances associated with CVE-2026-8452. The rule specifically monitors for HTTP GET requests targeting PHP files within the /var/vpn/theme/ directory, accompanied by efforts to execute shell commands (bin/sh) via the web server.
This rule detects network traffic patterns associated with the ARToken PhaaS backend infrastructure. It identifies specific fixed URI paths and cross-affiliate API route contracts used by the backend service, as well as known indicators of compromise for affiliate infrastructure, to monitor for potential credential harvesting or adversary-controlled service interaction.
Detects HTTP POST requests directed at common Machine Learning inference endpoints (e.g., /predict, /classify, /infer) where the request body claims to be a multipart form-data image upload (Content-Type: image/jpeg) but the actual binary data does not contain the standard JPEG file signature (FF D8 FF). This pattern is indicative of adversarial attempts to bypass security controls or provide malformed payloads to ML models.
Detects unauthenticated GET requests to known AI and machine learning infrastructure management endpoints (e.g., DCGM, Ray, Kubeflow) that are commonly exposed without authentication. These interfaces provide operational metrics, cluster status, or job management details, which can be leveraged for reconnaissance.
Detects unauthorized GET requests to common default paths associated with AI agent developer consoles and infrastructure, such as '/playground', '/api/predict', or '/_stcore'. These requests are identified by the absence of expected 'Authorization' or 'Cookie' headers, suggesting an unauthenticated probe or discovery attempt by an external actor looking for misconfigured or exposed AI development interfaces.
Detects attempts to exploit CVE-2026-48710 in Starlette frameworks where an attacker attempts to smuggle path or authority information via the HTTP Host header using character encoding variations like slashes or URL-encoded sequences.
Detects unauthenticated attempts to access Adobe Commerce/Magento REST API endpoints. This rule specifically looks for HTTP GET requests containing /rest/ and /V1/ paths without an Authorization header, which may indicate an attempt to exploit CVE-2026-71362.
This rule detects attempts to exploit a SQL injection vulnerability in Sangoma Switchvox (CVE-2026-9586) by inspecting HTTP POST request bodies and URIs for common SQL injection patterns such as UNION SELECT, SLEEP(), and DROP TABLE. This vulnerability allows an attacker to execute arbitrary SQL commands on the backend database.
This rule detects HTTP POST requests to the Apache Qpid Broker-J Management API that include a pre-set JSESSIONID in the cookie header during a login operation, indicating an attempt to exploit CVE-2026-92609 via session fixation.
Detects attempts to access administrative or configuration endpoints on Cisco Identity Services Engine (ISE) without providing valid authentication headers, cookies, or authorization tokens. This behavior is indicative of exploitation attempts targeting CVE-2026-20192 to gain unauthorized administrative access.
Detects incoming HTTP POST requests that match the signature of the China Chopper web shell client, specifically looking for common evaluative functions (eval, assert, base64_decode) being passed to single-letter parameters, which is a hallmark of this specific web shell's communication pattern.
Detects outbound HTTP POST requests characteristic of Emotet malware command-and-control beaconing. The rule identifies communication to an external network using a specific URI pattern (4-16 alphanumeric characters), a 'Content-Type: application/octet-stream' header, and a legacy Internet Explorer 7.0 User-Agent, while explicitly excluding common non-binary or HTML response formats in the request body.
Detects common SQL injection patterns (such as UNION SELECT, tautologies, sleep functions, and information_schema queries) within HTTP URI parameters and HTTP POST request bodies, which are indicative of an attempt to manipulate backend database queries.
Detects HTTP traffic patterns characteristic of Qakbot (QBot) malware command-and-control (C2) communication. This rule monitors for POST requests to .php files featuring a specific hardcoded User-Agent (IE 7.0 on Windows NT 5.1), a 'Cookie' header, 'application/octet-stream' content type, and a request body length of at least 201 characters.
Page 23 of 202