The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local shell scripts located at ~/scripts/create_worktree.sh and ~/scripts/cleanup_worktree.sh for repository management.
[COMMAND_EXECUTION]: It dynamically assembles humanlayer launch command strings using context (such as ticket IDs and feature names) extracted from external plan files.
[COMMAND_EXECUTION]: A critical security checkpoint is implemented in Step 3, which instructs the agent to present all details and the final command strings to the user for manual verification and approval before any execution occurs.

create-worktree