Skip to content
SkillModel-led2,390 tokens in use

Humanize Scan

A quick scout for Korean text: checks six AI tells that held up in testing, shows before and after counts, and says whether a full pass is worth it.

6,692 stars on the repository MIT Licenseepoko77-ai/im-not-ai · skills/humanize-scan
/plugin marketplace add epoko77-ai/im-not-ai /plugin install humanize-korean@im-not-ai
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing Humanize Scan costs

238tokens idle
2,390tokens 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
  • Structure16/25
  • Examples3/15
  • Output0/15
  • Size11/15
  • 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 →
Excellent
  • Development activity96
  • Maintenance100
  • Community64
  • Maturity28
  • Momentum40

Humanize Scan answers one question about a Korean text: is it worth running the full Humanize Korean pipeline? It looks at six signals only, rewrites the sentences that trigger them, and counts the signals before and after.

The whole job happens inside the chat. It writes no files, runs no scripts and starts no sub-agents. It returns a verdict line, a table of the six signals, three to five changed sentences and a closing that depends on how many signals fired.

What You Get

  • A verdict line saying whether little or a lot needs fixing
  • A table with each of the six signals before and after
  • Three to five before and after sentences of up to 100 characters each
  • The fixed full text when fewer than three signals fired

Common Use Cases

  • Deciding whether a Korean draft needs the full humanize run
  • Getting a fast look at an AI-written paragraph in chat
  • Showing a colleague which AI tells a text contains

What Makes It Unique

It limits itself to the six patterns that stayed above human levels across models and tasks, and it leaves out patterns that testing rejected, such as sentence-initial connectives and quotation marks. It refuses to give out a full text when many signals fired, because only six of 85 patterns were checked and a partial result could pass as finished work.

Under The Hood

The skill is one instruction file that reads the pasted text, or a file path you give it, and works entirely in the conversation. It lists the six patterns with their measured ratios, the things it never changes, such as facts, quotes and register, and the output format. It runs no scripts, starts no sub-agents, writes no files and does not fetch web pages or call outside APIs.

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