Skip to content
SkillMixed3,874 tokens in use

linkedin-humanizer

Strip the AI tells readers and LinkedIn's slop filter react to, audit a draft before publishing and learn your voice from a few of your own posts.

6,078 stars on the repository MIT Licensesergebulaev/linkedin-skills · skills/linkedin-humanizer
npx skills add sergebulaev/linkedin-skills
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing linkedin-humanizer costs

92tokens idle
3,874tokens in use
35,406tokens, 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 Mixed

Script-ledMixedModel-led

Scripts do part of the work and the model's judgement does the rest.

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

Touches

shelldetectednetworkdetectedwrites filesdetectedcredentialsdetected

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
  • Frontmatter11/15
  • Structure11/25
  • Examples9/15
  • Output10/15
  • Size13/15
  • Resources14/15

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 →
Fair
  • Runs shell commands−5

    does this on the machine, and the frontmatter does not say so

    SKILL.md, line 41
  • File that was not read−15

    a file the skill ships that was not read, so what it does is unknown

    scripts/detectors.env.example
  • Reaches the network−10

    does this on the machine, and the frontmatter does not say so

    scripts/test_detectors.py, line 32
  • Reads credentials−15

    does this on the machine, and the frontmatter does not say so

    scripts/test_detectors.py, line 58
  • Writes files−8

    does this on the machine, and the frontmatter does not say so

    scripts/test_detectors.py, line 357

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Excellent
  • Development activity96
  • Maintenance100
  • Community64
  • Maturity28
  • Momentum40

linkedin-humanizer rewrites LinkedIn text to remove patterns that read as machine-written: dense 2026 AI vocabulary, reveal bridges, stacks of short fragments, stacked triads and performed sincerity. Em dashes are capped by density, not banned.

It has two extra modes. Audit mode reviews a draft against a 20-point checklist and returns blockers and warnings without rewriting. Profile mode learns your voice from three to six of your own posts so the other writing skills match it. It does not promise to beat AI detectors.

What You Get

  • Rewritten text with a diff of what changed and why
  • Tell density per paragraph and a reader-read estimate of human, mixed or AI
  • An audit report with blockers, warnings and suggested fixes for a finished draft
  • A voice profile file the other LinkedIn skills read before drafting

Common Use Cases

  • Cleaning an AI-drafted post before it goes live
  • Running a pre-publish audit on a finished draft
  • Building a voice profile from a handful of past posts
  • Defending a stylistic choice such as an em dash by looking up the rule’s evidence
  • Showing how widely hosted AI detectors disagree on the same text

What Makes It Unique

It states its limits: no edit reliably beats a trained detector, so it reports the tells it found and fixed and never a detector score. It scrubs proportionally, so a clean draft gets two or three touches, and it checks its own output for the fingerprints humanizers leave behind, such as staged variance, framed confessions and a text with zero em dashes. It also never adds hedges or sincerity markers the author did not write.

Under The Hood

Most of the skill is instructions and reference files for rewriting, auditing and profile building, which run in the conversation. Profile mode writes the voice profile markdown file in the repository, and can read your recent activity from Apify when a token is set. One bundled Python script, used by the detector tester sub-skill, runs as a shell command. It reads detector API keys from the environment or a .env file, sends your draft text over the network to hosted AI detection services such as GPTZero, Originality.ai, ZeroGPT, Sapling and Copyleaks, and can write a JSON report file. It calls only the services you hold keys for and makes no network calls in its offline demo mode. The skill warns that the draft is disclosed to those providers. The mode examples in the skill file are usage notes for the skill, and the script is the only code it runs.

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