lark-doc

Installation

Summary

Create, read, edit, and manage Lark documents with content extraction, summarization, and formatting support.

Supports three document types: Docx, Wiki, and Cloud Documents; handles embedded spreadsheets, multidimensional tables, and whiteboards by extracting tokens for downstream skill routing
Core operations include fetch (with scoping and detail levels), create (XML or Markdown), update (str_replace, block operations, append, overwrite), and media management (insert, download, preview)
All docs +create, docs +fetch, and docs +update commands require --api-version v2 flag; defaults to DocxXML format unless Markdown is explicitly requested
Recognizes Feishu document URLs and tokens from both feishu.cn and doubao.com domains; routes based on URL path patterns and token type, not domain alone

SKILL.md

docs

场景与 Shortcut 路由

CRITICAL：先判断场景，再读取该场景的参考文件；不要在任务开始时一次性读取全部参考文件。每个文件只在首次进入对应阶段时读取一次。

身份：文档操作推荐显式指定 --as user。例外：如果 doc_token / note_doc_token 等是从 bot 链路（如 vc +detail --as bot → note +detail --as bot）取得的，应继续显式使用 --as bot，不要无条件切回 user——身份延续规则见 ../lark-shared/SKILL.md。

所有表示本地文件的 @path 均使用 @./xxx 形式的相对路径，并以运行 lark-cli 时的当前工作目录（CWD）为基准。