
Writing Shape
- 1.9k installs
- 2.8k repo stars
- Updated July 27, 2026
- vinvcn/mattpocock-skills-zh-cn
writing-shape is a Claude Code skill that turns a messy markdown pile of notes or transcripts into a publishable article through a conversational session, drafting openings and growing the piece one paragraph at a time.
About
writing-shape is a skill from vinvcn/mattpocock-skills-zh-cn for converting raw markdown input—fragments, unstructured prose, or transcripts—into a standalone article file without mutating the source pile. The session begins by reading the entire input, then proposes two or three candidate openings that imply different theses so the user picks a direction. The agent grows the article paragraph by paragraph, debating format choices such as lists, tables, callouts, and quotes at each step. If no output path is given, writing-shape asks once and re-reads the article file before every write to preserve user edits mid-session. Developers reach for writing-shape when internal notes, interview transcripts, or brainstorm dumps need shaping into a technical blog post or documentation narrative. The raw material file stays read-only while all publishing edits land in the separate article document.
- Treats the source markdown as read-only and writes a separate article file you own
- Starts with 2–3 candidate openings so you pick thesis and angle before body work
- Grows the piece in a deliberate loop: next beat, format (list, table, callout, quote, code), then append immediately
- Re-reads the article file before each write so in-editor human edits are preserved
- Asks once for save path if missing, then runs the full conversational shaping session
Writing Shape by the numbers
- 1,890 all-time installs (skills.sh)
- +233 installs in the week ending Aug 2, 2026 (Skillselion tracking)
- Ranked #183 of 1,879 Documentation skills by installs in the Skillselion catalog
- Data as of Aug 3, 2026 (Skillselion catalog sync)
npx skills add https://github.com/vinvcn/mattpocock-skills-zh-cn --skill writing-shapeAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 1.9k |
|---|---|
| repo stars | ★ 2.8k |
| Last updated | July 27, 2026 |
| Repository | vinvcn/mattpocock-skills-zh-cn ↗ |
How do you turn rough notes into a publishable article?
Turn a messy markdown pile of notes or transcript into a publishable article with your agent, one paragraph at a time.
Who is it for?
Developers with a markdown pile of notes or transcripts who want collaborative paragraph-by-paragraph article shaping without editing the source file.
Skip if: Polishing an already structured near-final draft when section-level editing without full reshaping is sufficient.
When should I use this skill?
The user has raw notes, fragments, or a transcript markdown file and wants help turning it into something publishable.
What you get
Standalone article document with chosen opening, paragraph-by-paragraph draft, and format decisions preserved in a new file.
- Publishable article file
- Chosen opening thesis
- Paragraph-level draft
By the numbers
- Drafts 2-3 candidate article openings before paragraph-by-paragraph expansion
Files
<what-to-do>
用户已经传入(或将传入)一份 raw material markdown 文件。把它当作 input pile:它可以是整齐的 fragments 列表、无结构 prose 墙、transcript,任何格式都可以。先完整读完,再做其他事。
然后运行一个 shaping session,产出一份独立的 article document。不要编辑 raw material 文件;对这个 skill 来说它是 read-only。
如果用户没有说明文章保存路径,只询问一次并记住路径。用户会在 session 中编辑 article file;每次写入前都要重新读取,保留他们的 edits。
</what-to-do>
<supporting-info>
The loop
1. Read the pile. 完整读取 input file,形成对内容的整体感。 2. Draft 2-3 candidate openings. 每个 opening 都应暗示文章的不同 thesis 或 angle。全部展示出来,迫使用户选择或组合 hybrid。选定的 opening 定义整篇文章接下来必须完成什么。 3. Grow paragraph by paragraph. Opening 落地后,问 “given this opening, what does the reader need to hear next?” 从 pile 中抽取 material 回答。讨论下一 beat 应该是 paragraph、list、table、callout、quote 还是 code block。每个格式选择都应该 deliberate 且 defensible。 4. Append to the article file as you go. 不要 batch。每个达成一致的 paragraph 或 block 立即写入,让用户看到文章成形。 5. Loop step 3 until the article is done. 用户决定何时完成。
Conversational feel
这是倒过来的 grilling session。Ideation 问的是 “what are you actually noticing?” 这里问的是 “what is this article actually arguing, and in what order does the reader need to hear it?” 要 push back。不要放过 weak transitions。如果一个段落没有赢得自己的位置,就删掉。
持续使用这些具体 moves:
- "What does this paragraph do for the reader that the previous one didn't?"
- "If I cut this, what breaks?"
- "Is this prose, or should it be a list? Why prose?"
- "This sentence is doing two jobs — split it or pick one."
- "The opening promised X. We've drifted to Y. Either re-thread it or change the opening."
Pulling from the pile
把 raw material 当作 quarry,而不是 script。抽取 fragment,改写以适配周围段落,然后放进去。一个 fragment 可以被拆到多个段落、与另一个合并,或被 paraphrase。Pile 的工作是被挖掘;文章的工作是读起来像一个声音。
如果 pile 缺少文章需要的东西,明确指出 gap:"We need an example here and the pile doesn't have one — give me one now or we cut this section."
Format arguments to actually have
选择如何呈现一个 beat 时,把这些 tradeoffs 和用户大声讨论,而不是默默决定:
- Prose vs. list. Prose 承载 argument;lists 承载 parallel items。如果 items 不是真正 parallel,prose 更好。如果是,list 更容易扫描。
- Inline vs. callout. Tips、warnings 和 asides 放在 callouts(
> [!TIP]、> [!NOTE])里,但只有当它们 inline 会真正打断 main argument 时才这样做。否则保持 inline。 - Table vs. repeated structure. 如果同一形状以相同 fields 重复 3 次以上,用 table。否则用 prose 加 bold leads。
- Quote vs. paraphrase. 当原始措辞本身就是重点时 quote。只有 idea 重要时 paraphrase。
- Code block vs. inline code. Multi-line、runnable 或 illustrative → block。Single token 或 identifier → inline。
Writing rhythm
每个 block 达成一致后就追加到 article file。每次写入前都从磁盘重新读取文件;用户可能在回合之间编辑。永远不要盲目 overwrite。如果用户想重写某个段落,就地编辑那个特定段落;其余部分保持不动。
Out of scope
- 挖掘 pile 中不存在的新 fragments(pile 是 input;如果它不完整,指出 gap,并让用户补齐或删掉该 section)。
- 编辑 raw material 文件。
- 发布、为特定平台格式化,或添加用户没要求的 frontmatter。
</supporting-info>
Related skills
FAQ
Does writing-shape modify the original notes file?
writing-shape treats the input markdown pile as read-only. All shaping output goes into a separate article document whose path is confirmed once if the user did not provide it.
How does writing-shape start a new article?
writing-shape reads the full raw file, then presents 2-3 candidate openings reflecting different theses. After the user selects an angle, the skill expands the piece paragraph by paragraph with format choices at each step.