Detections
Explore public detection logic contributed by the community across SIEM and rule languages.
5,879 detections
Filters
Last updated
All Time
Detection languages
2,954
1,378
313
289
239
Contributors
1,807
1,529
1,304
180
173
Categories
20,460
11,847
5,879
5,158
4,866
Platforms
3,747
1,857
1,076
782
160
Products / Services
2,241
1,434
1,273
746
582
MITRE Techniques
2,448
2,151
1,129
937
790
CVEs
19
15
9
7
5
IDS Classtypes
834
334
94
23
20
IDS Protocols
689
276
235
117
23
Detects network connections to the npm registry (registry.npmjs.org) initiated by AI coding agent processes (such as Claude, Cursor, Copilot, Windsurf, Codeium, Aider) or by unexpected processes that are not standard package managers or browsers.
Detects execution of a Python script named SysMon.py via a scheduled task, correlated with a subsequent network connection to ipinfo.io initiated by the same Python interpreter within a 10-minute window.
Detects endpoint network connections and correlates multi-stage communications associated with the ClickFix browser cache smuggling campaign, including initial payload retrieval and secondary stage/C2 communication across devices.
Detects DNS requests to major generative AI and Large Language Model API endpoints (such as OpenAI, Anthropic, Google Gemini, Mistral, Groq, DeepSeek, and others) initiated by processes other than standard web browsers (e.g., chrome, msedge, firefox, safari). This behavior may indicate unauthorized data exfiltration, shadow AI usage, or automated tooling/scripts interacting with LLM APIs.
Detects indicators of compromise including Bash-based reverse shells communicating with a C2 domain on port 7070, outbound network connections from web server and shell processes to the same C2 infrastructure, and the dropping of known suspicious plugin archive files (akismetv4-plugin.zip, Hlx.zip, template.zip).
Detects potential C2 beaconing activity associated with Potassium by correlating a DNS query for a known C2 domain with a subsequent TCP connection to port 50792 from the same device within a short time window.
Detects script-driven (curl, wget, python) outbound connections to
AI/LLM API endpoints. The campaign uses legitimate AI platform APIs as covert
command, control, and exfiltration channels, blending malicious traffic with
approved web services to evade network-based detection.
AI/LLM API endpoints. The campaign uses legitimate AI platform APIs as covert
command, control, and exfiltration channels, blending malicious traffic with
approved web services to evade network-based detection.
Detects execution and network activity associated with the cache_base operator toolkit, including JuicyPotatoNG (kitanai.exe), RegPwn (nandemo.exe), ufw.exe reverse shell, and upd-ms.bat launching these tools with connections to *.strangled.net on port 7070.
Detects sandbox or training hosts performing DNS resolution followed shortly by outbound connections to public IP addresses on ports 80 or 443, which may indicate a firewall misconfiguration or unauthorized egress from isolated environments.
Detects the execution of unapproved Remote Monitoring and Management (RMM) software tools followed by subsequent outbound public IP network activity from the same process on the same device, indicating an active remote session.
Detects a sequence where a device exhibits Remote Monitoring and Management (RMM) tool execution followed by high-volume outbound network connections to consumer cloud storage services (such as OneDrive or Google Drive) from the same device, potentially indicating data exfiltration via RMM.
Exposure-assessment hunt for endpoints running OpenCode and contacting OpenCode Console or AI provider APIs, to prioritise API key rotation after the October 2026 OpenCode breach.
Detects browser processes accessing a chatgpt.com/share/ link followed shortly thereafter by connections to multiple distinct external hosts, potentially indicating exfiltration or unauthorized interaction following an AI chat reference.
Detects browser processes accessing a chatgpt.com/share/ link followed shortly thereafter by connections to multiple distinct external hosts, potentially indicating exfiltration or unauthorized interaction following an AI chat reference.
Detects web server worker processes (such as php-fpm, apache2, nginx, w3wp.exe) spawning download utilities like curl or wget, certutil with transfer flags, or initiating /dev/tcp network connections, which may indicate web application exploitation or webshell activity.
Detects outbound network connections from processes targeting a known hard-coded Potassium campaign C2 IP address (134.199.154.64), followed by fallback DNS resolver queries using standard public DNS servers by the same process.
Detects non-browser processes contacting multiple distinct cloud storage APIs within a one-hour timeframe, a pattern indicative of potential multi-service command and control or data exfiltration.
Detects endpoint, network, and security log activity matching known indicators of compromise (IOCs) associated with PEERSTITCH / CL-STA-0954, including specific file hashes, IP addresses, and URLs.
Correlates an inbound connection to a Python/Xoscar/Xinference pool process with an outbound connection from a shell spawned by that same pool process shortly afterward, potentially indicating exploitation and reverse shell activity.
Detects a behavioral sequence where a signed process loads an unsigned DLL from a user-writable path (DLL sideloading) followed shortly by outbound HTTPS network connections from the same process to legitimate cloud storage or API infrastructure.
Detects a process with no prior AI API activity in the past 14 days suddenly generating a high volume (200+ calls) of network connections to public AI inference endpoints (OpenAI, Anthropic, Grok, OpenRouter). This behavior is consistent with stolen API keys being used to drain resources or exfiltrate data from a newly compromised host.
Page 2 of 280