The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is designed to generate and execute complex shell and Python scripts for system-level automation.
[COMMAND_EXECUTION]: Instructions provide templates for creating persistent background tasks using systemd timers and cron jobs, which involve modifying system configuration and maintaining execution across sessions.
[COMMAND_EXECUTION]: The debugging section explicitly recommends using sudo -u user for manual verification of scripts, which facilitates privilege escalation within the environment.
[DATA_EXFILTRATION]: The skill mentions using network-capable tools such as curl for Slack notifications and sendmail for email alerts, creating potential pathways for data to leave the system.
[PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it translates user-provided task requirements into executable code.
Ingestion points: User-provided task descriptions and automation requirements processed according to the instructions in SKILL.md.
Boundary markers: None identified; the skill lacks specific delimiters or instructions to disregard embedded commands in the processed data.
Capability inventory: The agent is equipped with Bash, Python, Write, Edit, systemctl, cron, curl, and sudo capabilities.
Sanitization: There are no documented procedures for sanitizing or validating user-provided task inputs before they are incorporated into generated scripts.

agent-automation-scripter