The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the deepline CLI to perform status checks and catalog searches. Evidence includes instructions to run deepline --maintenance off org status and deepline -h. These are documented as read-only operations used for environment verification and planning.
[SAFE]: The skill implements strong 'Human-in-the-Loop' controls. It explicitly forbids executing paid or state-changing calls, installing skills, or starting recurring work without a single confirmation step that covers both the plan and the verified organization name.
[SAFE]: The skill uses the --maintenance off flag for all pre-confirmation CLI checks to ensure that no automatic updates or background installations occur before the user has approved the plan.
[SAFE]: The skill includes logic to prevent 'org spoofing' by requiring a live identity check (org status) immediately before handing off to an implementation skill, ensuring the execution environment matches the user's explicit approval.

deepline-start