The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell scripts to perform security inspections and verify port availability on the target server. Evidence: The inspect.sh script executes network and system diagnostic tools such as ss, netstat, and sshd -T to report the server's security posture. The verify.sh script uses Bash's /dev/tcp device to test for locally open ports; while this triggered a reverse shell heuristic, it is used here as a benign connectivity check.
[PRIVILEGE_ESCALATION]: The skill requires administrative access to modify system configurations and manage security services. Evidence: The references provided in the references/ directory use sudo to perform tasks such as editing /etc/ssh/sshd_config and managing firewall rules with ufw or firewall-cmd.
[INDIRECT_PROMPT_INJECTION]: The skill processes inspection data from a remote server, which constitutes a potential injection surface if the server is already compromised. 1. Ingestion points: Output from scripts/inspect.sh. 2. Boundary markers: The skill instructs the agent to summarize results for the user. 3. Capability inventory: Shell command execution via SSH and sudo file writes. 4. Sanitization: No explicit sanitization of command output is performed.

server-hardening