The Agent Skills Directory

[DATA_EXFILTRATION]: The script scripts/agent.py accesses several sensitive system files to perform integrity checks.
The check_file_integrity function reads files such as /etc/shadow, /etc/sudoers, and /etc/ssh/sshd_config to compute SHA-256 hashes.
While these checks are restricted to computing hashes and are standard for File Integrity Monitoring (FIM) tools, the act of reading these files involves processing sensitive system data.
Additionally, the script allows bypassing TLS verification via an environment variable, which could expose API communications to interception.
[COMMAND_EXECUTION]: The script scripts/agent.py executes system commands via the subprocess module.
The run_osquery_check function executes the osqueryi binary to gather host information.
Currently, the query used is hardcoded to list root processes, which is a benign administrative task.
[EXTERNAL_DOWNLOADS]: The SKILL.md file provides instructions to download and install security software from official sources.
Fetches the Wazuh GPG key and agent package from the official packages.wazuh.com repository.
Downloads the Windows MSI installer for Wazuh from the same official domain.
[PROMPT_INJECTION]: The skill processes external data which creates a surface for indirect prompt injection.
Ingestion points: scripts/process.py parses alert files, and scripts/agent.py consumes data from the Wazuh API.
Boundary markers: None identified in the processing logic.
Capability inventory: The skill can execute subprocesses (osqueryi) and read system files.
Sanitization: No explicit validation or sanitization of external data before use in logs or outputs was observed.

configuring-host-based-intrusion-detection