Quirón skill
A skill for Claude Code, Codex and Cursor: strips signs of AI writing and checks the result against human-writing ranges.
- 66 ★ on the repository
- Last commit
- AGPL-3.0
Extended access
Why this level
- The skill rewrites text and the result must not be passed off as human writing where readers have a right to know the author.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/quironDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The skill rewrites text and the result must not be passed off as human writing where readers have a right to know the author.
This is third-party code. Review the repository files before installing.
What it does
The skill edits text against a list of 33 signs of machine writing: triples, extra headings, summary sections, "not just X, it's Y" constructions and long dashes. After the edit the `aimeter.py` script measures 23 rates and compares each with the range found in human writing from before ChatGPT. The model does not grade itself: the meter flags both overshoot into choppy plainness and leftover machine style. The edit uses only facts from your text and notes, and `scripts/factdiff.py` lists numbers and links absent from the source. Registers are blog, fiction and Spanish; other languages get the checklist without a meter. Scripts use the Python 3 standard library and no API key is needed.
Who it is for. Authors of internal docs, READMEs and drafts who want text without machine tells and a measurable check of the edit.
Good fit when
- A README or how-to sounds like a ChatGPT reply and needs tidying.
- An email draft is checked for AI tells without rewriting yet.
- The edit must be confirmed not to have made the text too choppy.
- An author brings notes and a dry draft is rewritten from them.
Not a fit when
- Machine text must be passed off as human on social media, in articles, schoolwork, reviews or applications: the author explicitly forbids this.
- Passing detectors like GPTZero and Turnitin is required: the skill does not beat them.
- The text is under about 120 words: the meter does not run on it.
Example request
Make docs/launch.md sound less machine-written using my notes below and show the meter result before and after the edit.Limitations
The skill is meant for texts you write for yourself or your team. The author forbids passing the edit off as human writing in public or deceiving readers. It does not beat detectors that read token probabilities. Languages beyond English and Spanish have no human baseline: the meter says so and shows only numbers like paragraph length and heading count. Texts under about 120 words are not measured. It was tested on text from three Claude models and one GPT model.
How to disable. Run `npx skills remove quiron` or delete the skill folder from the agent directory. Remove the Claude Code plugin with `/plugin uninstall quiron@quiron`.
FAQ
Does it beat AI text detectors?
No. Detectors like GPTZero and Turnitin read token probabilities and the skill does not touch them. It works on what a human reader notices.
Does it make things up?
The instructions forbid it: an edit uses only facts from your text and notes. `scripts/factdiff.py` lists numbers and links absent from the source. When detail is missing the skill asks.
Which scripts exist?
`python3 scripts/aimeter.py FILE` measures 23 rates, `scripts/audit.py --brief FILE` prints the pattern checklist and `scripts/check.sh FILE` runs one pass of the full loop.
Which languages does the meter support?
English (blog and fiction) and Spanish. Other languages get the checklist without a human baseline.
Is an API key needed?
No. The skill runs on the model you already use, and the scripts need only the Python 3 standard library.
Discussion
No questions or reviews yet.
Related
SkillsA skill that finds AI writing tells, rewrites the text in a chosen voice, and optionally scores it from 0 to 100
Two skills for English prose: humanize rewrites a model draft in a natural voice, and ai-check flags machine-writing tells
A de-AI writing skill: it repairs the narrative architecture of fiction and edits professional documents by their venue rules
Humanizer (brandonwise)
humanizer
A skill and Node.js CLI that score English text for AI-writing patterns and statistics and suggest fixes on a 0 to 100 scale
An editor for AI-written Russian text: removes bureaucratese, cliches and 38 AI patterns while keeping facts intact
im-not-ai
im-not-ai (humanize-korean)
A Claude Code plugin skill that rewrites AI-written Korean text into a natural human style without touching the meaning
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| Quirón (this card) | no | 66 | |
| Humanizer | no | 68,355 | gives the model a list of AI writing signs to remove; Quirón adds a script that measures 23 rates against human ranges |
| Stop Slop | no | 18,896 | asks the model to grade itself on five questions; Quirón measures the text with a script |
README in short
The README explains how the skill differs from other humanizers: measurement instead of a model's self-grade. It separately covers responsible use, the pattern list, the meter and example prompts.