Skip to content
SkillModel-led4,702 tokens in use

work-humanizer

Edit AI-sounding work writing so a careful person could have written it, keeping every fact, or interview the writer first to get real specifics.

680 stars on the repository MIT Licensephuryn/product-engineering · work-humanizer/skills/work-humanizer
claude plugin install work-humanizer@product-engineering
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing work-humanizer costs

35tokens idle
4,702tokens in use
5,540tokens, 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. Scripts: check.py.

Touches

shelldetected

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 →
Good
  • Frontmatter8/15
  • Structure19/25
  • Examples3/15
  • Output10/15
  • Size15/15
  • Resources15/15

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 activity100
  • Maintenance100
  • Community44
  • Maturity4
  • Momentum40

work-humanizer edits text that reads like a model wrote it: posts, emails, updates, docs and announcements. It keeps every fact and invents nothing, and it works in two modes. Rewrite mode edits text you hand over. Interview mode asks up to five short questions when the text could have come from anyone with the same one-line request.

The skill is built on three measurements: readers notice texture, detectors read composition and not words, and specifics carry the human signal. It never promises text that detectors will miss.

What You Get

  • The edited text with every number, name, date, commitment and caveat unchanged
  • For interview mode, up to five numbered questions followed by the final text
  • Optional notes of up to five bullets on what changed, with spots where a real specific would help
  • A shorter version on request, with the cut named

Common Use Cases

  • Making a Slack update, support reply or release note sound written by a person
  • Turning notes or a voice memo into a post, newsletter or pitch email in the writer’s own voice
  • Cleaning a landing page of marketing filler
  • Checking a draft against a list of AI tells before it goes out

What Makes It Unique

The skill treats the problem as composition and not vocabulary. It cites a detector test in which rewording AI drafts stayed flagged while text built from the writer’s own sentences scored human, so its answer is an interview that gathers the writer’s material. It lists tells by strength (staging, doing the reader’s work, inflation, misplaced hedges, rhythm, punctuation, chat residue), and bans adding facts.

Under The Hood

The skill is a SKILL.md of instructions with two supporting files: tells.md, lists of AI vocabulary and filler for the final sweep, and check.py, a small Python script that reads one text file or stdin and prints counts of dashes, semicolons, stock words, filler transitions and sentence-length bands. The script uses only the standard library, makes no network calls and writes nothing. The agent runs it only if it can run code. Otherwise the skill reads and edits text 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