Detections
Explore public detection logic contributed by the community across SIEM and rule languages.
2,832 detections
Filters
Last updated
All Time
Detection languages
1,375
495
363
281
175
Contributors
848
453
393
252
188
Categories
20,453
11,837
5,876
5,154
4,865
2,832
Platforms
1,449
916
693
670
264
Products / Services
994
701
613
469
216
MITRE Techniques
1,240
912
484
482
388
CVEs
6
5
2
1
1
IDS Classtypes
15
10
6
5
3
IDS Protocols
39
Detects anomalous, high-volume querying against machine learning inference endpoints. The rule identifies potential model extraction or data exfiltration attempts by flagging high counts of requests from a single user/API key, particularly during off-hours or with high variance in unique inputs, which may indicate automated efforts to reconstruct proprietary models or extract sensitive training data.
Detects anomalous, high-volume querying against machine learning inference endpoints. The rule identifies potential model extraction or data exfiltration attempts by flagging high counts of requests from a single user/API key, particularly during off-hours or with high variance in unique inputs, which may indicate automated efforts to reconstruct proprietary models or extract sensitive training data.
Detects the installation of Model Context Protocol (MCP) servers or AI agent plugins that exhibit characteristics of supply chain risk, such as unverified or unknown publishers, requests for highly sensitive permissions (filesystem write, credential read, unrestricted network egress), or recent public listing, which aligns with AI model supply chain poisoning techniques (AML.T0104).
Detects suspicious AI inference API activity indicative of credential compromise or misuse. The rule identifies three main signals: rapid authentication failures (credential stuffing), anomalous geolocation/ASN usage for successful requests, and successful API calls occurring after a key rotation event.
Detects instances where AI agents are activated through external communication channels like webhooks, email, or chat, and the provided payload includes potentially malicious imperative instructions (e.g., 'delete', 'execute', 'send credentials'). The detection logic excludes activity originating from trusted, allow-listed internal sources to reduce false positives.
Detects adversarial attempts to infer whether specific records were used in an AI model's training set by analyzing raw prediction confidence scores and identifying suspicious query patterns such as near-duplicate submissions or systematic sweeps.
This rule detects high volumes of near-identical, incrementally-varied queries directed at an AI model API. This behavior is indicative of a black-box adversarial optimization loop, where an attacker attempts to infer model vulnerabilities, bypass safety filters, or perform model evasion by observing how slight input variations change model responses.
This rule detects high volumes of near-identical, incrementally-varied queries directed at an AI model API. This behavior is indicative of a black-box adversarial optimization loop, where an attacker attempts to infer model vulnerabilities, bypass safety filters, or perform model evasion by observing how slight input variations change model responses.
This rule detects high volumes of near-identical, incrementally-varied queries directed at an AI model API. This behavior is indicative of a black-box adversarial optimization loop, where an attacker attempts to infer model vulnerabilities, bypass safety filters, or perform model evasion by observing how slight input variations change model responses.
This rule detects high volumes of near-identical, incrementally-varied queries directed at an AI model API. This behavior is indicative of a black-box adversarial optimization loop, where an attacker attempts to infer model vulnerabilities, bypass safety filters, or perform model evasion by observing how slight input variations change model responses.
This rule detects high volumes of near-identical, incrementally-varied queries directed at an AI model API. This behavior is indicative of a black-box adversarial optimization loop, where an attacker attempts to infer model vulnerabilities, bypass safety filters, or perform model evasion by observing how slight input variations change model responses.
This rule detects high volumes of near-identical, incrementally-varied queries directed at an AI model API. This behavior is indicative of a black-box adversarial optimization loop, where an attacker attempts to infer model vulnerabilities, bypass safety filters, or perform model evasion by observing how slight input variations change model responses.
This rule detects anomalous behavior in AI agent tool invocations, specifically monitoring for high volumes of tool calls, high costs associated with token consumption, or suspicious tool interactions following a flagged indirect prompt injection. This behavior is indicative of potential resource exhaustion (denial of service) or abusive agent behavior as defined by the MITRE ATLAS framework for AI security.
This rule detects anomalous behavior in AI agent tool invocations, specifically monitoring for high volumes of tool calls, high costs associated with token consumption, or suspicious tool interactions following a flagged indirect prompt injection. This behavior is indicative of potential resource exhaustion (denial of service) or abusive agent behavior as defined by the MITRE ATLAS framework for AI security.
Detects Microsoft Entra ID sign-in events where a user authenticates from a known residential proxy ASN, matching the user's historical geography but using a new, unrecognized device fingerprint. This behavior is indicative of session token replay attacks, where an adversary uses stolen session cookies to bypass standard authentication.
Detects instances where a Google Cloud Config Connector (KCC) service account performs a SetIamPolicy operation that grants high-privilege roles (roles/owner or roles/resourcemanager.organizationAdmin) at the organization or folder scope. This activity often indicates an out-of-band configuration change, a compromised CI/CD pipeline, or an unauthorized escalation, as legitimate KCC changes should be tracked via GitOps records.
Detects unauthorized or suspicious use of Google Cloud Config Connector (KCC) service accounts to apply high-privilege IAM roles ('roles/owner' or 'roles/resourcemanager.organizationAdmin') at the Organization or Folder scope. This rule monitors for 'SetIamPolicy' API calls where the principal is identified as a KCC-related service account, flagging potential exploitation of KCC's authority to grant excessive permissions.
Detects unauthorized or suspicious use of Google Cloud Config Connector (KCC) service accounts to apply high-privilege IAM roles ('roles/owner' or 'roles/resourcemanager.organizationAdmin') at the Organization or Folder scope. This rule monitors for 'SetIamPolicy' API calls where the principal is identified as a KCC-related service account, flagging potential exploitation of KCC's authority to grant excessive permissions.
This rule detects anomalous, high-volume activity from a single user account interacting with known public generative AI platforms (e.g., ChatGPT, Copilot, Claude). It uses a threshold-based approach on CloudAppEvents logs to identify potential reconnaissance or bulk information gathering activities which could signify unauthorized use or automated data scraping via AI interfaces.
Detects instances where a single user account exhibits an anomalously high frequency of requests to public generative-AI platforms within a one-hour window, potentially indicating automated reconnaissance or data staging activities using AI tools.
Detects GitHub repository actions (git.clone, oauth_authorization.create, repo.download_zip) from actor IPs not seen for that user in the prior 30 days, requiring either 3+ actions or activity across 2+ distinct repositories from the new IP within the same day. Aggregating to burst/multi-repo behavior (rather than alerting on a single new-IP event) filters out one-off false positives from travel, VPNs, or dynamic IP reassignment while still surfacing sustained anomalous access patterns consistent with account compromise or bulk exfiltration.
Page 31 of 135