What problem does it solve? Handling Lark (Feishu) approvals manually through the client is slow when you need to search launchable approval definitions, review pending tasks, or act on approval instances at scale. This Skill lets an AI agent perform the full approval workflow through lark-cli commands. ## Core Features & Use Cases - Approval Task Processing: Query pending/completed approval tasks and approve, reject, transfer, roll back, add-sign, or remind on them using instance_code and task_id pairs. - Approval Initiation: Search launchable approval definitions, inspect form controls and node lists, then create native approval instances with correctly structured form values. - Instance Management: View instance details, cancel initiated instances, query initiated instances by definition, and add CC recipients. - Use Case: Ask the agent to "approve my pending leave requests" — it queries the todo list, retrieves instance_code and task_id, and executes the approval with a comment. ## Quick Start Ask the assistant to list your pending Lark approvals and approve a specific one, or to search for a leave approval definition and submit a new instance.