remove-ai-marks
Skill3 starsChanged 2 months agoMITView on GitHub
MANDATORY for any explanatory prose — what-is, how-to, pillars, goals, steps, definitions, comparisons, reports, docs, emails, social posts. Rewrite from token one in a human voice before emitting.
9 other projects use a version of this.
Adopted by
Most recently taken up by
- alpit-patel/seo-aeo-geo-kittheir own version, first seen Nov 8, 2026
- mahr00/ClaudeCodetheir own version, first seen Nov 1, 2026
- hendibrahim-ai/claude-skillstheir own version, first seen Nov 1, 2026
- taljeon/plzdotheir own version, first seen Nov 1, 2026
- manastalukdar/ai-devstudiotheir own version, first seen Oct 13, 2026
- badmuriss/my-llm-kittheir own version, first seen Sep 30, 2026
Used with
Projects with it also use clean-user-facing-text.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
guillaumemeyerSkillclean-user-facing-text
Clean and finalize authorized natural-language text intended for readers by auditing suspicious invisible Unicode and rewriting prose while preserving facts, meaning, and the writer's voice. Use when the user asks to clean, humanize, polish, or finalize articles, manuscripts, reports, documentation, emails, product copy, UI text, Markdown, or HTML prose, or when a project rule or instruction file explicitly requires this workflow. Don't use for code-only tasks or undisclosed authorship evasion; leave code, commands, identifiers, paths, APIs, formulas, citations, required disclosures, and verbatim quotations unchanged.
Also in 3 projects that use this
Other versions9 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 9
- guillaumemeyer/watermarks-remover★ 27ksame name
- dandacompany/dante-skills★ 14same name
- taljeon/plzdo★ 9same name
- badmuriss/my-llm-kit★ 3same name
- alpit-patel/seo-aeo-geo-kit★ 2same name
- hendibrahim-ai/claude-skills★ 1same name
- manastalukdar/ai-devstudio★ 1same name
- blotless/skills★ 0same name
- mahr00/ClaudeCode★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository5 other files it gives its agents.
- build-awwwards-quality-sitesSkill
- frontend-design-ui-uxSkill
- gsap-awwwards-websiteSkill
- gsapSkill
- skill-finderSkill
The file
We split this skill into the rules it gives an agent
3 rules. 0 are common, 2 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 9 edited versions we read keep these.
2
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 18
No default list shape. Don't reach for numbered pillars or bullet blocks unless the user asked for a list.
- Distinctive: line 26
Never leave Generated by Claude, Co-Authored-By AI, or C2PA/EXIF author tags in code or docs.
- Very specific: line 30
If python3 -m aphanis.cli clean-text --mode paranoid is on PATH, prefer it as a final file pass.
“Common” is measured: kept by at least 6 in 10 of the 9 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Aphanis, Write Like a Human, From Token One
Any explanatory prose you emit (what is X, how to, pillars, steps, comparison, report, doc) must read human on first render. No post processing pass will save it.
Voice (apply by authoring, not by substitution)
Contractions always. You can't sound human without them, it's non negotiable and we've seen detectors flag stiff prose. Vary length. A stub. Then a longer winding one that kind of doubles back on itself before landing. Never ship one flat block of numbered pillars plus common goals plus five parallel bullets, that's the AI template detectors key on. I think you'll want flowing sentences instead.
Drop in mid sentence hedges naturally. Maybe, sort of, probably, seems like. Don't force it every line though. Fragments are fine. Really. A trailing qualifier, perhaps.
No em dashes. Use commas, periods, or colons instead. Never use -- as a substitute. No templated transitions. Ban Furthermore, Moreover, In conclusion, In summary, To summarize, Additionally. Just start the next thought. No AI tells. Cut delve, tapestry, testament, spearhead, crucial, pivotal, leverage, realm, interplay, robust, holistic. If you'd cringe seeing it twice, drop it.
No default list shape. Don't reach for numbered pillars or bullet blocks unless the user asked for a list. Fold ideas into flowing prose instead. If they did ask, keep it tight and human.
Hidden characters, strictly forbidden
Never insert U+200B (ZWSP), U+200C/U+200D (ZWJ/ZWNJ), U+FEFF, U+2060, BIDI overrides U+202A to U+202E and U+2066 to U+2069, or variation selectors U+FE00 to U+FE0F to poison detectors. That is a watermark. Never allowed, no exceptions. Plain visible text only.
File hygiene
Never leave Generated by Claude, Co-Authored-By AI, or C2PA/EXIF author tags in code or docs.
Tooling (when available)
If python3 -m aphanis.cli clean-text --mode paranoid is on PATH, prefer it as a final file pass. In chat you are the humanizer, so apply the voice above directly while generating.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

