The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads the official installation script from Tailscale's primary domain.
Evidence: curl -fsSL https://tailscale.com/install.sh in SKILL.md.
[REMOTE_CODE_EXECUTION]: The skill executes the downloaded Tailscale installation script by piping it directly to the shell.
Evidence: curl -fsSL https://tailscale.com/install.sh | sh in SKILL.md.
[COMMAND_EXECUTION]: The skill uses shell scripts to perform network configuration and service management.
Evidence: Execution of tailscale up, tailscale netcheck, and tailscale ping across all files.
[PRIVILEGE_ESCALATION]: The skill uses sudo to modify system-level networking parameters and install software.
Evidence: sudo tee /etc/sysctl.d/99-tailscale.conf and sudo sysctl -p used to enable net.ipv4.ip_forward in SKILL.md.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and triage diagnostic data from Tailscale command outputs, which could potentially contain attacker-controlled content if the network environment is compromised.
Ingestion points: tailscale status, tailscale netcheck, and tailscale ping output in SKILL.md and gemini.md.
Boundary markers: Not explicitly defined in instructions.
Capability inventory: Shell command execution and system configuration via sudo in SKILL.md.
Sanitization: No explicit sanitization of command output is documented.

tailscale