What problem does it solve?
It helps convert scattered notes, fragments, or rough prose into a cohesive, reader-ready article by actively shaping structure, flow, and format decisions.
Core Features & Use Cases
- Guided shaping session: Reads an input markdown pile end-to-end and produces a separate article document without editing the source.
- Thesis-first openings: Drafts 2–3 candidate openings that imply different angles, then aligns the rest of the article to the chosen opening.
- Paragraph-by-paragraph growth: Iteratively decides what comes next and argues for the best representation (prose, lists, tables, callouts, quotes, or code blocks) based on reader needs.
- Preserves user edits: Re-reads the target article file before writing each new block so user modifications aren’t lost.
- Gap detection: Identifies missing examples or content the pile doesn’t contain and requests the user to fill them or cuts weak sections.
Quick Start
Use writing-shape with your rough notes file, and ask the agent to shape it into a publishable article by selecting an opening and growing the piece block-by-block into a dedicated article document.