feishu-fetch-doc

The feishu-fetch-doc skill enables AI agents to seamlessly integrate with the Feishu/Lark ecosystem by retrieving cloud document content in a structured Markdown format. Designed for developers, researchers, and power users who rely on Feishu for documentation and collaboration, this tool bridges the gap between private enterprise knowledge bases and AI processing workflows. It automates the extraction of complex document structures, including text, tables, and internal block elements, while providing a specialized protocol for handling binary media such as images, files, and whiteboard sketches.

• Converts Feishu/Lark docx content into clean, Lark-flavored Markdown for LLM consumption.

• Fully compatible with Feishu Wiki (space node) navigation, automatically resolving object types (docx, sheet, bitable) before extraction.

• Integrates with the feishu_doc_media utility to download binary assets like images, attachments, and whiteboard snapshots using unique resource tokens.

• Supports direct URL or token-based identification for both standard documents and complex knowledge base (Wiki) entries.

• Provides a robust architectural bridge for multi-tool orchestration, allowing agents to fetch text and media in distinct, efficient steps.

• Users must provide the doc_id or full Feishu URL; the system automatically extracts necessary tokens for API communication.

• For Wiki links, always utilize the feishu_wiki_space_node tool first to determine the resource type (e.g., docx, sheet, or bitable) to avoid type mismatch errors.

• Images, files, and whiteboards are returned as HTML-style placeholders (e.g., <image token=.../>); these require a secondary call to the feishu_doc_media tool for local storage and retrieval.

• Ensure the environment has the appropriate Feishu integration permissions configured to access the specified document scopes.

• This tool is best used in sequence: first resolve the document type, then fetch the primary content, and finally trigger media downloads if the output contains media tags.