Detections
Explore public detection logic contributed by the community across SIEM and rule languages.
4,916 detections
Filters
Last updated
All Time
Detection languages
2,236
1,064
466
410
326
Contributors
725
665
535
485
271
Categories
1,085
870
805
765
576
Platforms
1,843
1,022
856
814
656
Products / Services
988
596
477
456
383
MITRE Techniques
18,115
15,635
12,577
8,388
6,215
4,916
CVEs
37
17
17
15
12
IDS Classtypes
196
73
47
29
16
IDS Protocols
301
27
24
22
6
Detects a single authenticated user account or client performing a high volume of requests to sequential object IDs on the same API path within a short window. This pattern is characteristic of Broken Object Level Authorization (BOLA) or Insecure Direct Object Reference (IDOR) exploitation, where an attacker attempts to gain unauthorized access to data records that do not belong to them by iterating through predictable numeric identifiers.
This rule detects attempts to exploit an unauthenticated administration API vulnerability (CVE-2026-35616) in FortiClient EMS versions 7.4.5 and 7.4.6, commonly referred to as FortiBleed. This vulnerability could allow an unauthenticated attacker to gain unauthorized access to the administration interface.
Detects potential zero-click LLM prompt injection attacks against Microsoft 365 Copilot. The rule correlates Microsoft 365 Copilot activity involving external/untrusted URLs or domains with recent inbound email activity to the same user. This pattern is designed to identify scenarios where an attacker leverages an inbound email to trigger a Copilot interaction with malicious external content, potentially leading to unauthorized data access or exfiltration.
Detects potential OAuth device code phishing attacks by identifying HTTP requests to the legitimate microsoft.com/devicelogin URL containing suspicious user_code or device_code parameters. This pattern indicates an adversary is using an OAuth device authorization grant flow to steal tokens or gain unauthorized access without the victim needing to provide credentials directly to the attacker.
This rule detects potential prompt injection attempts targeting AI systems by looking for specific keywords related to AI models (gpt, claude, copilot) in HTTP logs, combined with suspicious characters or commands commonly used in injection attacks (e.g., <script>, javascript:, eval(). The rule triggers if 5 or more such attempts are observed from a single IP within an hour.
This rule detects potential mass emailing or spam activity by identifying instances where a single email account sends more than 100 emails within a one-minute window. This behavior is often characteristic of compromised accounts performing phishing campaigns or bulk unauthorized communication.
This rule detects rapid sequential DNS queries by a single host for multiple domains that mirror common brand impersonation patterns (e.g., login, business, empresas, post-app). The detection logic flags when at least two unique domains matching these patterns are queried within an 18-minute window, which is indicative of potential credential harvesting or phishing reconnaissance.
This rule detects instances where a user clicks on a URL identified as potentially malicious (phishing-related) and subsequently has a risky sign-in event in Azure AD. The rule correlates URL click events from email with Azure AD sign-in logs, specifically looking for sign-ins marked as risky or originating from suspicious user agents (e.g., specific axios versions). The phishing URLs are identified by keywords such as 'azureapplicationregistration.pages.dev', 'chrnobinson.com', 'workers.dev', '.pages.dev', '.web.core.windows.net', '/oauth2/v2.0/authorize', 'client_id=', and 'scope=openid'.
Detects high-volume unauthenticated API queries against a ServiceNow instance by a Guest user. This behavior is indicative of an adversary attempting to exfiltrate sensitive information, such as ticket data, configuration secrets, or employee records, via the 'related_list_edit' REST endpoint or standard table APIs.
Detects connections to domains associated with the Montana Empire phishing kit, which leverages AI-hallucinated postal service domains for credential harvesting. This kit targets phantom domains identified through LLM hallucination prediction and exploits AI-generated domain suggestions.
Detects connections to AI-hallucinated API endpoints commonly generated by coding assistants. These phantom domains are weaponized by adversaries to intercept CI/CD pipeline telemetry, build notifications, and developer authentication tokens from automated systems.
Detects potentially malicious activity within Microsoft 365 involving the creation of inbox rules for message forwarding/redirection, modifications to mailbox forwarding configurations, or mass outbound email sending. These actions are common indicators of persistent access, email exfiltration, or automated reconnaissance after account compromise.
Detects a malicious HTA lure disguised as a military pay calculator. The attack involves a JScript routine that renames an embedded data file ('styles.dat') to an executable ('styles.dat.exe') and triggers its execution, characteristic of a spearphishing attachment delivering malware.
Detects potential exploitation attempts of CVE-2026-35616 targeting Fortinet FortiClient EMS versions 7.4.5 and 7.4.6. The rule identifies suspicious indicators such as API authentication bypass attempts marked by missing certificate headers, unauthorized modifications to remote access VPN profiles, and unauthenticated administrative API access on port 8013.
This rule detects potential 'impossible travel' scenarios within Azure AD sign-in logs. It identifies users who successfully sign in to Office 365 from two geographically distant IP addresses within a one-hour window, suggesting a potential compromise or an attempt to bypass security controls.
This rule detects attempts to exploit a Server-Side Request Forgery (SSRF) vulnerability in Oracle PeopleSoft by targeting the '/PSIGW/HttpListeningConnector' endpoint. The detection looks for malicious URI patterns and request bodies containing internal-only network indicators such as loopback addresses (127.0.0.1, localhost), link-local addresses (169.254.169.254), or local file access attempts (file://), which are indicative of SSRF exploitation attempts.
This rule detects the execution of Python (pip) or Node.js (npm) package managers with command-line arguments specifying alternative package repositories (e.g., --index-url, --registry) that do not belong to the standard, trusted public repositories (pypi.org or npmjs.com/npmjs.org). This behavior is often indicative of supply chain attacks, where attackers attempt to trick users or automated systems into installing malicious dependencies from compromised or attacker-controlled repositories.
Detects usage of pip or npm package managers where non-canonical registry or index URLs are explicitly specified. This behavior can be indicative of software supply chain poisoning, specifically attacks involving typo-squatting or AI-hallucinated malicious package repositories.
This rule detects a correlation between an unauthorized IAM user access event and the subsequent execution of cryptocurrency mining tools within the same AWS account within a one-hour window. This is a strong indicator of an adversary leveraging compromised cloud credentials to perform resource hijacking.
Detects suspicious process execution by GitHub Runner.Worker or Runner.Listener on CI runners. It specifically targets the execution of scripts or binaries associated with compromised action components (such as trivy-action or setup-trivy) where version tags have been tampered with to inject credential-harvesting code. The rule monitors for unauthorized command line arguments or file paths linked to these components.
This rule monitors for potential lateral phishing activity within an organization by identifying internal accounts sending high volumes of emails, correlating this with MailItemsAccessed audit events from unexpected IPs, and detecting OAuth token abuse patterns (such as granting consent or adding credentials) within the Exchange Online environment.
Page 23 of 235