The Agent Skills Directory

[SAFE]: The skill implementation follows a human-in-the-loop security model, explicitly instructing the agent to wait for user approval after proposing a plan and before editing any files in Phase 4.
[SAFE]: No unauthorized network operations or data exfiltration patterns were detected; the skill's functionality is limited to analyzing and modifying local files and user-provided images.
[SAFE]: The skill instructions emphasize adhering to existing project constraints, design tokens, and surrounding code style, which prevents the introduction of unverified external dependencies or configuration drift.

refactor-ui