The Agent Skills Directory

[SAFE]: The skill performs legitimate configuration management by installing agent templates in ~/.claude/agents/ or .claude/agents/. It explicitly instructs the agent to show diffs and ask for user permission before overwriting any existing files.
[SAFE]: The use of model: haiku in the sub-agent templates is a documented feature of the platform to scope model usage and does not represent a security risk.
[SAFE]: The skill uses $ARGUMENTS to determine whether to install globally or at the project level (--project), which is standard parameter handling for this environment.

limite