What problem does it solve? Working with Lark (Feishu) cloud storage programmatically is error-prone: different document types use different token formats, wiki links hide the real document token, and async operations like import, export, and folder moves require polling. This Skill gives an AI agent precise instructions to upload, download, move, comment on, and manage permissions for Drive files without token or API mistakes. ## Core Features & Use Cases - File & Folder Operations: Upload and download files, create folders, copy/move/delete items, and list folder contents via lark-cli drive shortcuts and native APIs. - Document Comments: Add full-document or text-selection (local) comments on docx documents, list and count comments/replies with correct semantics, and handle restrictions like resolved or whole-document comments. - Import/Export & Async Tasks: Convert local files (Word, Markdown, Excel, CSV) into cloud documents, export docx/sheet/bitable to PDF/XLSX/CSV/Markdown, and poll async task results with a unified +task_result shortcut. - Use Case: A user pastes a wiki link and asks to add a comment on a specific paragraph. The Skill first resolves the wiki token to the real docx token via wiki.spaces.get_node, then adds a local comment with drive +add-comment --selection-with-ellipsis. ## Quick Start Ask the agent to upload a local file to a Lark Drive folder or add a comment to a Feishu document URL, for example: upload report.pdf to my Lark cloud space folder and grant me admin access.