lark-doc

Read, create, and edit Lark Docs and Wiki documents via lark-cli.

Updated Sep 13, 2026
One-click install
npx skills add https://github.com/U1traVeno/skills --skill lark-doc-u1traveno
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: lark-doc
Source: https://github.com/U1traVeno/skills/tree/main/profiles/lark/lark-doc
Command: npx skills add https://github.com/U1traVeno/skills --skill lark-doc-u1traveno

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? Working with Lark (Feishu) cloud documents programmatically requires navigating complex APIs for reading, creating, and precisely editing Docx and Wiki content. This Skill wraps those operations into lark-cli shortcuts so an AI agent can fetch, create, update, and manage document media without hand-writing API calls. ## Core Features & Use Cases - Document Read/Write: Fetch documents with scoped reading (outline, section, range, keyword) and create or update content using XML or Markdown formats with block-level precision (str_replace, block_insert_after, block_replace, block_delete, block_move_after). - Media & Resource Management: Insert images from clipboard or file, preview and download document media and whiteboard thumbnails, and manage Docx cover images. - History & Statistics: List document history versions, revert to a specific version, and count total words/characters with a dedicated script. - Use Case: A user pastes a Feishu docx URL and asks to summarize one section, then append a new chapter with a Mermaid diagram—the Skill fetches the section by heading anchor, then inserts structured XML content at the right block. ## Quick Start Ask the AI to read the Lark document at your docx URL and summarize its outline, making sure lark-cli auth login has been completed first.

Frequently Asked Questions about lark-doc

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I read a Feishu document from a URL?▼

Use lark-cli docs +fetch with the document URL or token. Choose --detail simple for reading, with-ids for locating blocks, or full before editing, and use --scope outline, section, range, or keyword to read only the part you need.

How to edit a specific paragraph in a Lark doc?▼

First fetch with --detail with-ids to get block IDs, then run docs +update with commands like block_replace, block_insert_after, or block_delete targeting the block ID. For plain text changes, str_replace with --pattern works without block IDs.

Should I use XML or Markdown format for Lark docs?▼

XML is the default and preferred format, especially for precise block-level edits, since it carries block structure and styles. Use Markdown only when the user explicitly provides a .md file or asks for Markdown import.

Can this skill manage document comments or spreadsheet data?▼

No. Comment management is handled by the lark-drive skill, and embedded spreadsheet or bitable data operations are routed to lark-sheets or lark-base after extracting their tokens from the document content.

Why did my block update fail after a previous edit?▼

Block IDs can become invalid after overwrite, block_replace, or block_delete operations. Re-fetch the document with --detail with-ids to obtain fresh block IDs before continuing block-level edits.

How do I count words in a Feishu document?▼

Fetch the document with docs +fetch --detail full --format json and pipe it into the bundled doc_word_stat.py script. The script reports word_count and char_count using its own counting rules, which are the authoritative standard.