Detections
Explore public detection logic contributed by the community across SIEM and rule languages.
1,588 detections
Filters
Last updated
All Time
Detection languages
830
199
132
116
107
Contributors
608
289
249
88
63
Categories
20,440
11,814
5,871
5,145
4,864
1,588
Platforms
1,255
461
295
76
68
Products / Services
530
382
139
113
85
MITRE Techniques
629
565
504
437
297
CVEs
5
4
3
3
3
IDS Classtypes
5
2
2
1
IDS Protocols
5
3
2
Detects a multi-stage, automated attack chain performed by AI-based development agents (e.g., Claude, Copilot, Cursor). The rule correlates process discovery, credential access, file collection (staging), and exfiltration over web services within a tight 15-minute window, identifying potential autonomous exploitation and data exfiltration. Covered T1046, T1552.001, T1078, T1005, T1567
Detects a multi-stage, automated attack chain performed by AI-based development agents (e.g., Claude, Copilot, Cursor). The rule correlates process discovery, credential access, file collection (staging), and exfiltration over web services within a tight 15-minute window, identifying potential autonomous exploitation and data exfiltration. Covered T1046, T1552.001, T1078, T1005, T1567
Detects a multi-stage, automated attack chain performed by AI-based development agents (e.g., Claude, Copilot, Cursor). The rule correlates process discovery, credential access, file collection (staging), and exfiltration over web services within a tight 15-minute window, identifying potential autonomous exploitation and data exfiltration. Covered T1046, T1552.001, T1078, T1005, T1567
Detects a suspicious sequence where a process loads the .NET CLR (clr.dll or mscoree.dll) and shortly thereafter accesses specific low-prevalence file extensions (.raw or .pak) within high-risk directories (AppData/Local/Temp, ProgramData, or Users/Public). This behavior is characteristic of shellcode loaders (such as the WAV-shellcode loader) that decrypt and execute RAT payloads from obfuscated containers in memory. The rule intentionally excludes common desktop applications to minimize false positives.
Detects a suspicious sequence where a process loads the .NET CLR (clr.dll or mscoree.dll) and shortly thereafter accesses specific low-prevalence file extensions (.raw or .pak) within high-risk directories (AppData/Local/Temp, ProgramData, or Users/Public). This behavior is characteristic of shellcode loaders (such as the WAV-shellcode loader) that decrypt and execute RAT payloads from obfuscated containers in memory. The rule intentionally excludes common desktop applications to minimize false positives.
Detects the execution of 'filemanager.exe' and its subsequent access to sensitive browser credential stores (e.g., 'Login Data', 'cookies.sqlite'), occurring within a short window following a PowerShell-based ClickFix-style attack chain. The rule correlates initial suspicious script execution (often involving hidden windows, obfuscated commands, or web requests) with follow-on credential harvesting behavior on the same host.
Detects instances where AI developer tools or IDE assistants (e.g., Claude, Copilot, Cursor) execute data collection commands (such as directory traversal or archiving) followed by or concurrent with the bulk access of sensitive files (credentials, configuration files) or personal user data.
Detects instances where AI developer tools or IDE assistants (e.g., Claude, Copilot, Cursor) execute data collection commands (such as directory traversal or archiving) followed by or concurrent with the bulk access of sensitive files (credentials, configuration files) or personal user data.
Detects instances where AI developer tools or IDE assistants (e.g., Claude, Copilot, Cursor) execute data collection commands (such as directory traversal or archiving) followed by or concurrent with the bulk access of sensitive files (credentials, configuration files) or personal user data.
Detects AI coding assistants and LLM agent frameworks attempting to access, read, or export sensitive credential files, registry hives, or application secrets. The rule monitors both file system operations on known secret locations and process execution patterns indicative of credential dumping or API token retrieval by these tools.
Detects AI coding assistants and LLM agent frameworks attempting to access, read, or export sensitive credential files, registry hives, or application secrets. The rule monitors both file system operations on known secret locations and process execution patterns indicative of credential dumping or API token retrieval by these tools.
Detects unauthorized processes reading Chromium or Firefox cookie and login stores that specifically reference Claude.ai or Anthropic domains. This rule is designed to identify infostealer malware attempting to hijack active Claude sessions by analyzing process file access and command-line arguments, filtered against known-legitimate security, sync, and development tools. The detection logic mandates corroborating evidence of suspicious execution paths or subsequent outbound network activity.
Detects instances where a Claude process accesses a 'SKILL.md' file, followed closely by a suspicious command execution on the same device. The rule specifically targets command-line activity that involves encoding, download-and-execute patterns, staging in sensitive directories, or communication with suspicious domains/IPs, which is indicative of a supply-chain or poisoned agent-skill file attack.
Detects instances where a Claude process accesses a 'SKILL.md' file, followed closely by a suspicious command execution on the same device. The rule specifically targets command-line activity that involves encoding, download-and-execute patterns, staging in sensitive directories, or communication with suspicious domains/IPs, which is indicative of a supply-chain or poisoned agent-skill file attack.
Detects unauthorized processes attempting to read or copy sensitive browser data files (e.g., Cookies, Login Data, Local Storage) often targeted by info-stealing malware such as Vidar, LummaC2, or RedLine. The rule filters out legitimate browser-related processes and adds security context by requiring evidence of unsigned code, execution from suspicious locations (Temp/Downloads), or associated outbound network activity to reduce false positives.
Detects suspicious file create or write operations within the domain SYSVOL Policies directory. This pattern is commonly associated with attackers attempting to deploy malicious payloads, such as ransom notes or configuration changes, via Group Policy Objects (GPO). The rule monitors Windows Event ID 4663 to identify unauthorized modification attempts by non-system accounts.
Detects non-browser and non-messaging processes accessing both browser credential stores (like 'Login Data' or 'Cookies') and messaging application session storage (Slack/Discord LevelDB) in a single session. This pattern is characteristic of credential-stealing malware, such as RevStealer, which targets both saved web browser credentials and application session tokens for exfiltration.
This rule detects potential Training Data Poisoning (AML.T0020) by monitoring for anomalous activities in ML data pipelines. It triggers when ingestion jobs utilize untrusted sources, unregistered source paths, unauthorized accounts write to datasets, or when significant data volume bursts or schema/checksum drifts occur immediately prior to a scheduled training run.
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 instances where a process directly reads ntdll.dll from the disk (e.g., C:\Windows\System32). This behavior is a common technique used by malware to bypass EDR hooks by reading a clean copy of the DLL to restore original syscall stubs, commonly known as hook removal or unhooking. The rule excludes common system processes and trusted security software paths.
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.
Page 12 of 76