operating-tailscale-agent-hosts
Pass
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill provides numerous shell commands for system configuration and operation, including a
readiness-check.shscript that performs read-only system information gathering (hostname, kernel version, memory, disk space, and service status). - [EXTERNAL_DOWNLOADS]: The documentation references the installation of standard, well-known tools (Tailscale, GitHub CLI, Docker, tmux, pnpm) via official distribution package managers (e.g.,
pacman). These references target trusted services and legitimate software repositories. - [CREDENTIALS_UNSAFE]: The skill demonstrates safe credential handling practices. In
references/setup.md, it instructs the user to capture sensitive input (Wi-Fi passwords) usingread -rspto avoid echoing secrets to the terminal or logs. It explicitly warns against printing pre-shared keys (PSKs) or tokens in output visible to the agent. - [DATA_EXFILTRATION]: The instructions emphasize selective configuration transfer, advising users to exclude credentials, histories, caches, and machine-specific trust records by default when moving data between hosts.
- [PRIVILEGE_ESCALATION]: Administrative commands using
sudoare included for legitimate host setup tasks, such as modifying network configurations and managing systemd units. These operations are transparently documented and consistent with the skill's primary purpose of host administration.
Audit Metadata