Skip to content
SkillModel-led1,340 tokens in use

Writing Shape

Read a pile of raw notes and shape it into an article paragraph by paragraph, choosing the opening, grounding each concept and arguing over format.

328,791 stars on the repository MIT Licensemattpocock/skills · skills/in-progress/writing-shape
claude plugin install mattpocock-skills@claude-plugins-official
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing Writing Shape costs

19tokens idle
1,340tokens in use

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

Nothing detected: no shell, network, file writes, git or credentials in its files.

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Fair
  • Frontmatter8/15
  • Structure19/25
  • Examples6/15
  • Output10/15
  • Size15/15
  • Resources0/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent

No findings.

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Excellent
  • Development activity96
  • Maintenance100
  • Community56
  • Maturity32
  • Momentum40

Writing Shape is the exploit half of a writing workflow. It reads your raw material file end to end, treats it as read-only, and runs a shaping session that produces a separate article document.

It drafts two or three candidate openings with different theses, then grows the piece block by block, asking what the reader needs to hear next.

What You Get

  • A separate article file that grows as each block is agreed
  • Several candidate openings, each implying a different thesis
  • A running list of grounded concepts
  • Explicit format arguments for prose versus list, callout versus inline, and quote versus paraphrase
  • Named gaps when the pile lacks an example

Common Use Cases

  • Turning a transcript into a structured post
  • Shaping a wall of unstructured notes into an argument
  • Tightening a draft by asking what a paragraph does for the reader
  • Finding gaps in source material before publishing

What Makes It Unique

It runs as a grilling session inverted: the agent pushes back on weak transitions, asks what breaks if a paragraph is cut, and names missing material aloud. Format choices are argued in the open with fixed tradeoffs for each pair, and the raw file is never edited.

Under The Hood

The skill is a single Markdown file plus an agent metadata file for Codex. It reads the raw material file in full, writes each agreed block to the article path, and re-reads that file before every write. It runs no scripts and reaches no network service. It is user-invoked only.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers