The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a wide range of shell commands to install, configure, and verify network services. This includes package installation via apt, service management via systemctl, and network enumeration using tools like nmap, snmpwalk, smbclient, and enum4linux.
[PRIVILEGE_ESCALATION]: Extensive use of sudo is required for the skill's primary functions, including updating package lists, installing software, modifying configuration files in /etc/, and updating firewall rules (UFW/Windows Firewall).
[DATA_EXFILTRATION]: The skill includes patterns for manual data harvesting from system logs. It specifically instructs the agent to search for "POST" requests in web server logs to extract potential credentials, which represents a data exposure surface.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external sources (logs and tool outputs), creating a surface for indirect prompt injection if those sources are controlled by an attacker.
Ingestion points: Reading Apache access/error logs (tail -f) and parsing outputs from enumeration tools like smbclient or snmpwalk.
Boundary markers: The skill implements a strong "Mandatory confirmation gate" requiring the user to confirm targets, scope, and specific commands before execution.
Capability inventory: The skill has high capabilities, including file system writes, service configuration, network probing, and execution of shell commands with root privileges via sudo.
Sanitization: There is no automated sanitization of external data; safety relies on the manual confirmation gate and human review of command effects.

network-101