Open-source AI Security Operations Center: alert fusion, LLM-agent triage, MITRE ATT&CK investigation, and a replayable decision ledger for every agent step. Self-hostable, runs with no API keys, MIT licensed. Ships an MCP server for Claude, Cursor and Continue.
Python agentic-aiai-securityalert-triageblue-teamcybersecuritydetection-engineeringincident-responselanggraphllm-agentsmcpmitre-attackmodel-context-protocolpurple-teamsecurity-operationsself-hostedsiemsoarsocthreat-detectionthreat-intelligence
id: det-endpoint-303
name: Kernel Module Load Or Unload Via auditd Watch
description: |
Triggers when the AiSOC auditd profile records init_module / finit_module /
delete_module syscalls or execution of /sbin/insmod, /sbin/rmmod, /sbin/modprobe.
Pivots on auditd_key so the rule fires regardless of how the kernel module
primitive was invoked. Loadable kernel modules are the standard delivery
vehicle for Linux rootkits (Diamorphine, Reptile, Symbiote, etc.); legitimate
module loads are dominated by boot-time systemd-modules-load.service and
filesystem drivers and should be easy to allowlist by parent process.
version: 1.0.0
severity: medium
tags:
- mitre.attack.t1547.006
- mitre.attack.t1014
- tlp.white
category: endpoint
log_source:
product: linux
service: auditd
detection:
condition: |
auditd_key == "aisoc_priv_esc_module_load"
AND exe NOT ENDSWITH_ANY ["/systemd-modules-load", "/kmod"]
false_positives:
- Boot-time module load by systemd-modules-load.service
- Driver hot-plug (USB device, NVIDIA installer, ZFS/DKMS rebuild)
- Container runtime loading overlay/br_netfilter on host
playbook: tpl-defense-evasion
enabled: false
author: AiSOC
created: '2026-05-12'
modified: '2026-05-12'
quarantine_reason: "authored as YAML with no Python spec, so the engine never loaded it; no connector is known to emit its fields, so translating it would add a rule that cannot fire"