The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses SSH to interact with a remote NAS and executes local shell scripts for mount management. These operations are core to the skill's purpose of host-specific management and troubleshooting.
[PRIVILEGE_ESCALATION]: The instructions direct the agent to use sudo for specific tasks such as mounting filesystems, installing LaunchDaemons, and modifying network settings. While these are high-privilege actions, they are explicitly scoped to the skill's primary function of infrastructure maintenance.
[PERSISTENCE]: The skill manages service persistence through launchd and /Library/LaunchDaemons/ configuration files to ensure NAS mounts are maintained across system reboots.
[INDIRECT_PROMPT_INJECTION]: The skill processes output from remote commands (such as dmesg logs and storage status) which serves as an ingestion point for potentially untrusted data.
Ingestion points: Output from SSH commands executed on the three-body NAS, as defined in SKILL.md.
Boundary markers: None present in the instructions to delimit command output from agent instructions.
Capability inventory: SSH command execution, local shell script execution, sudo access, and launchctl service management.
Sanitization: None specified for the interpolated command outputs.

three-body