The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the subprocess and os modules to manage system networking configurations through various command-line utilities.
The NetworkSecurityMonitor class in network-security.md executes iptables commands via subprocess.run to dynamically block IP addresses based on detected activity.
The RouteManager and BGPNeighborManager classes in routing.md execute route, ip, and vtysh commands to manipulate routing tables and BGP peering sessions.
The WireGuardManager in vpn.md utilizes subprocess to interface with wg and wg-quick for tunnel management.
[PRIVILEGE_ESCALATION]: Several scripts included in the skill explicitly request or utilize elevated privileges to perform restricted networking tasks.
The run_certbot utility in tls-ssl.md is designed to automatically prepend sudo to commands when necessary for managing SSL/TLS certificates.
The VLAN management shell script in subnetting.md uses sudo to execute ip link, ip addr, and ip route commands required for Layer 2 and Layer 3 segmentation.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from multiple external sources, creating a potential surface for indirect injection.
Ingestion points: Raw network packet bytes are parsed manually in dns.md and network-security.md. Additionally, the output of external system commands is parsed in routing.md and vpn.md to determine current system state.
Boundary markers: The provided code examples lack explicit boundary markers or strict delimiters when processing these external data streams.
Capability inventory: The skill has access to powerful capabilities including raw socket access, system command execution, and file system modifications.
Sanitization: The code examples, particularly run_certbot in tls-ssl.md, utilize potentially unsafe patterns like shell=True with string-joined arguments, which could be exploited if input parameters like domain names or email addresses are not rigorously validated.

networking