Document to Mind Map

Problem

Long documents, meeting notes, web pages, papers, and scanned images often contain dense information, but readers usually need a reusable structure rather than a restatement of the source. Manual summarization can easily produce confused hierarchy, missing key points, overly long leaf nodes, or poor structure choices. This skill is aimed at that workflow: it compresses source text into headings and short nodes while preserving the original document skeleton, making it useful for engineers, product managers, and researchers who need a discussable knowledge frame.

How it works

The skill first determines whether the user wants to understand content, extract a framework, or generate a specific diagram. When attachments are provided, it prioritizes reading or recognizing the text and avoids fabricating unreadable content. It then preserves heading hierarchy, condenses final-level paragraphs into one-sentence nodes, and removes noise and duplication. Output is Markdown: # for the root theme, ## and ### for sections and knowledge points, and bullet lists for details. It maps content to a structure: mind_free for divergent summaries, mind_right for process steps, mind_org for organization charts, mind_ishikawa_left for root-cause analysis, mind_timeline_h for milestones, and mind_treeTable_left_title for comparisons or parameter lists. After cloud submission, it returns complete Markdown, an online view link, and an image link, with later edits regenerating the full result.

Boundaries

This skill works best when the source has a clear text skeleton or readable recognized text. It may be limited by low-quality scans, purely visual content, or unreadable attachments. It prioritizes faithful transformation and does not invent facts to fill gaps. Levels beyond Markdown heading limits become lists. The generated map is intended for review and editing, so users should still verify key conclusions, data, and citations.