Detections
Explore public detection logic contributed by the community across SIEM and rule languages.
960 detections
Filters
Last updated
All Time
Detection languages
215
211
197
168
63
Contributors
283
142
113
112
85
Categories
670
218
168
112
95
Platforms
39,915
7,245
6,529
3,864
3,627
960
Products / Services
713
68
64
59
59
MITRE Techniques
310
305
146
136
134
CVEs
5
3
3
2
1
IDS Classtypes
13
11
11
9
8
IDS Protocols
62
1
Detects outbound HTTP/HTTPS POST requests where the URI, query string, or path contains evidence of sensitive credential material, such as AWS access keys, private key blocks, or common environment variable patterns, indicative of potential credential harvesting and exfiltration to an external destination.
Detects a sequence where an AWS Secrets Manager secret is retrieved via CloudTrail, followed within 30 minutes by an EDR-monitored process initiating an outbound network connection on a non-standard port (not 80 or 443). This behavior is consistent with an adversary using stolen credentials to pivot into a persistent command-and-control channel.
This rule detects potential cost harvesting or unauthorized resource abuse by monitoring API endpoints (typically AI model gateways) for high-volume or resource-intensive queries associated with a specific identity. It correlates these patterns with billing or autoscaling trigger events (e.g., cost spikes, quota exhaustion) and identifies agentic sessions that exhibit abnormally long reasoning loops, which suggests an attempt to waste computational budget.
Detects anomalous, high-frequency enumeration of AI/ML assets such as model registries, container image repositories, and training data buckets. This activity is indicative of reconnaissance for AI supply chain attacks, specifically tracking with MITRE ATLAS AML.T0007 (Discover AI Artifacts), where an actor probes infrastructure to identify targets for model theft, data poisoning, or unauthorized access.
Detects anomalous, unauthorized access to machine learning model artifacts (e.g., .onnx, .pt, .h5) within object storage or model registries. The rule identifies suspicious patterns including high-volume file reads by non-ML team identities, lack of multi-factor authentication, or access from geolocations outside approved operating regions, indicative of potential model exfiltration or intellectual property theft.
Detects an anomalous volume of artificial intelligence model inference API calls (e.g., AWS Bedrock, Azure OpenAI, Google Vertex AI) initiated by a single IAM principal or API key within a short time window. This activity is indicative of 'LLMjacking', where compromised cloud credentials are leveraged by threat actors to perform unauthorized, high-cost AI model inference tasks.
Detects a two-stage activity: first, the potential unauthorized reading or access of LLM provider API keys from files, environment variables, or command-line arguments; and second, the subsequent use of those same credentials to authenticate to LLM provider APIs (OpenAI, Anthropic, Azure, Bedrock) from a different, unrecognized source IP address, indicating potential credential theft and session hijacking.
Detects a two-stage activity: first, the potential unauthorized reading or access of LLM provider API keys from files, environment variables, or command-line arguments; and second, the subsequent use of those same credentials to authenticate to LLM provider APIs (OpenAI, Anthropic, Azure, Bedrock) from a different, unrecognized source IP address, indicating potential credential theft and session hijacking.
Detects anomalous, high-velocity enumeration of IAM entities and cloud infrastructure resources by a single principal within a 5-minute window. This behavior is consistent with automated reconnaissance performed by agentic AI or scripted tools, often used to map cloud environments prior to exploitation.
Detects when a namespace-scoped identity creates, updates, or patches AWS or Azure IAM-related custom resources in Kubernetes. This pattern monitors for a 'confused deputy' scenario where a user might attempt to leverage high-privilege cloud operators (like ACK or ASO v2) that operate with a shared high-privilege identity to modify cloud IAM configurations.
Detects when a namespace-scoped identity creates, updates, or patches AWS or Azure IAM-related custom resources in Kubernetes. This pattern monitors for a 'confused deputy' scenario where a user might attempt to leverage high-privilege cloud operators (like ACK or ASO v2) that operate with a shared high-privilege identity to modify cloud IAM configurations.
Detects the use of common command-line tools (curl, wget, python, etc.) to interact with known cloud provider metadata service IP addresses or URL paths, often performed by shell processes. This behavior is frequently associated with initial reconnaissance or attempts to steal identity tokens from cloud-based virtual machines.
Detects bulk read operations on AWS Secrets Manager or similar cloud secret stores where the accessed resources contain keywords related to cryptocurrency custody, private keys, or wallet material. This pattern is indicative of unauthorized exfiltration attempts by threat actors such as Slim Spider targeting digital asset infrastructure.
This rule detects potential cloud service enumeration activity originating from a specific, potentially malicious IP address (23.234.84.102). It flags instances where this source IP performs specific API calls ('GetCallerIdentity' and 'ListTopics') without expected user-agent strings typical of legitimate AWS service traffic.
Detects the TeamFiltration/UNK_CondorFiltration spray signature: a large
number of distinct target UserPrincipalNames within a single tenant, each
receiving only a handful of sign-in attempts, spread across many distinct
AWS EC2-hosted source IPs/regions within a short time window. Individual
sign-ins from AWS-hosted IP ranges (VPN gateways, corporate cloud proxies,
remote contractors) are not flagged on their own; the rule requires both
the account-fanout and the IP-rotation dimensions to co-occur for the same
tenant in the same window.
number of distinct target UserPrincipalNames within a single tenant, each
receiving only a handful of sign-in attempts, spread across many distinct
AWS EC2-hosted source IPs/regions within a short time window. Individual
sign-ins from AWS-hosted IP ranges (VPN gateways, corporate cloud proxies,
remote contractors) are not flagged on their own; the rule requires both
the account-fanout and the IP-rotation dimensions to co-occur for the same
tenant in the same window.
Detects the TeamFiltration/UNK_CondorFiltration spray signature: a large
number of distinct target UserPrincipalNames within a single tenant, each
receiving only a handful of sign-in attempts, spread across many distinct
AWS EC2-hosted source IPs/regions within a short time window. Individual
sign-ins from AWS-hosted IP ranges (VPN gateways, corporate cloud proxies,
remote contractors) are not flagged on their own; the rule requires both
the account-fanout and the IP-rotation dimensions to co-occur for the same
tenant in the same window.
number of distinct target UserPrincipalNames within a single tenant, each
receiving only a handful of sign-in attempts, spread across many distinct
AWS EC2-hosted source IPs/regions within a short time window. Individual
sign-ins from AWS-hosted IP ranges (VPN gateways, corporate cloud proxies,
remote contractors) are not flagged on their own; the rule requires both
the account-fanout and the IP-rotation dimensions to co-occur for the same
tenant in the same window.
Detects a distributed password spraying campaign targeting cloud accounts, characterized by a large number of distinct accounts being targeted simultaneously from multiple AWS-hosted IP addresses, with a low volume of attempts per account per hour to evade standard threshold-based detection.
Detects a distributed password spraying campaign targeting cloud accounts, characterized by a large number of distinct accounts being targeted simultaneously from multiple AWS-hosted IP addresses, with a low volume of attempts per account per hour to evade standard threshold-based detection.
Detects a distributed password spraying campaign targeting cloud accounts, characterized by a large number of distinct accounts being targeted simultaneously from multiple AWS-hosted IP addresses, with a low volume of attempts per account per hour to evade standard threshold-based detection.
Detects a distributed password spraying campaign targeting cloud accounts, characterized by a large number of distinct accounts being targeted simultaneously from multiple AWS-hosted IP addresses, with a low volume of attempts per account per hour to evade standard threshold-based detection.
Detects high-volume, automated attempts to validate AWS credentials (access keys) from a single source IP address within a short time window. This pattern, characterized by a large number of distinct access keys being tested against identity-confirming API calls (e.g., GetCallerIdentity, AssumeRole) with a mix of success and failure, is indicative of automated credential harvesting or reconnaissance campaigns.
Page 12 of 46