Skip to content
SkillModel-led1,954 tokens in use

Better Writing

Writes and reviews interface copy such as labels, errors, empty states and confirmations, matching the product's voice and telling people what to do next.

12,497 stars on the repository MIT Licensejakubkrehel/skills · skills/better-writing
npx skills add jakubkrehel/skills
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing Better Writing costs

35tokens idle
1,954tokens in use
2,922tokens, up to

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Up to adds the files its instructions link. 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 →
Excellent
  • Frontmatter8/15
  • Structure25/25
  • Examples12/15
  • Output15/15
  • Size15/15
  • Resources10/15

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 →
Good
  • Development activity96
  • Maintenance52
  • Community48
  • Maturity24
  • Momentum40

Better Writing is an agent skill for interface copy. It starts by inventorying the strings a product already has, the nouns used for objects and the verbs used for actions, then writes or reviews new copy with those terms and one consistent voice.

It flags copy that misleads people or leaves them without a way forward, such as vague errors, generic link text and double-negative toggles. A deliberate brand voice is respected unless it creates ambiguity or translation risk.

What You Get

  • A term inventory built from translation calls and locale files, so new copy reuses the product’s own nouns, verbs and case policy
  • Rewrites of errors into instructions, such as a password rule stated as a requirement and a failed save with a next step
  • Destructive flow templates: a title naming the action and object, a body stating what is lost and buttons that pair a verb with an object
  • Empty-state patterns that say what the place is and offer one next action, with filtered variants that name the query
  • A detection table for strings such as Click here, Oops, Are you sure, successfully and concatenated plurals

Common Use Cases

  • Rewriting vague error messages across a signup form
  • Replacing a Yes and No confirmation dialog with a clear verb and object
  • Fixing toggles labeled with Don’t or Disable so they describe the on state
  • Replacing string concatenation and singular and plural ternaries with a message with plural arguments

What Makes It Unique

It prefers undo over confirmation for reversible, frequent actions and reserves typed confirmation for destroying accounts, workspaces or repositories. It treats taste as a non-finding: rewording that only suits the reviewer is out, while inconsistency, ambiguity and dead ends are in.

Under The Hood

The skill is a SKILL.md with a patterns reference and a reporting reference. It has no scripts. The agent searches your code for translation calls and locale files such as JSON, po and strings files, reads the surrounding copy and checks each label, error and term against the inventory. Strings that come from a server or CMS and cannot be seen are marked Not verified. A review leaves files untouched, and edits happen when you ask for fixes.

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