Skip to content
MichaelDoesMarketingPublic

About

Don't bite someone else's style. A Claude Skill that checks whether the idea you're about to publish already belongs to someone, and what you owe them. Sources, dates, and an honest coverage statement every time.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Latest commit

 

History

1 Commit

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 

Repository files navigation

sampled

Don't bite someone else's style.

Hip hop settled this decades ago. There's sampling, which is how most things get made and which requires credit. There's arriving somewhere on your own, which happens constantly and is nobody's fault. And there's biting, which is taking something, forgetting where you got it, and presenting it as yours.

Everybody outside that culture just carries vague anxiety about it and hopes nobody notices.

sampled runs on your own work before it goes out. You bring the framework, the thesis, the positioning line, the name, the concept. It tells you which of the three you're doing, component by component, with sources and dates.

The idea that makes it work

Resemblance doesn't decide it. Exposure does.

Two ideas can be nearly identical and both be honest, if neither person could have met the other. Two ideas can look different and one still be biting, if somebody absorbed the structure somewhere and rebuilt it in their own words.

So this isn't similarity matching. Most of the work is reconstructing what you were actually exposed to. What you read, who you follow, who they quote, what's canonical enough in your field that you absorbed it without noticing. Convergence is the default assumption, and biting is the claim that has to earn its way in with all three of close resemblance, established exposure, and clear priority.

The thing it usually finds is cryptomnesia. You read it, kept the structure, lost the citation, and later experienced it as your own thought. That's ordinary memory, not dishonesty, and it has ended careers of people who weren't lying.

The two rules that don't bend

It can never tell you something is original. It can tell you what it searched, what it found, and what it couldn't check. Every single output ends with its coverage stated, especially the clean ones, because a clean result that hides its blind spots is worse than no result. Books, podcasts, conference talks, private communities, and anything verbal are all effectively invisible to a search like this.

No claim without a source, a date, and a link. All three or it doesn't get said. An undocumented lineage claim about a named person is defamation wearing a research report.

What it won't do

It runs on your work. It's not a callout machine.

Checking a specific originality claim someone made publicly is fine, on the documented record only. Building a dossier on a person, assembling a pattern of their borrowing, or producing something meant to be posted as an accusation is out. The search coverage here isn't good enough to accuse anybody of anything, and a confident report built on partial coverage is how people get hurt.

Install

git clone https://github.com/MichaelDoesMarketing/sampled.git ~/.claude/skills/sampled

Or grab the packaged .skill from releases.

Try it with

"About to publish this framework. Has someone already done it?"

"I have a nagging feeling I read this somewhere and I can't place it."

"Is my name for this too close to something that exists?"

Real demo: I ran it on my own skill

The flagship in this portfolio is battle-tested, which applies battle rap mechanics to marketing copy. It's the thing I'm proudest of. So that's what it ran on.

THE CALL

Application: converging, and crowded. Structure: yours. Name: not claimable by anybody.

THE COMPONENTS

The application, battle rap mechanics applied to marketing and copywriting. Not new. Two direct neighbors:

Corey Patton, "What Battle Rap Has Taught Me About Communication", Medium, September 25 2025. Four lessons: performance, precision and the punch, character under pressure, cultural insight. Same source material, aimed at communication rather than marketing, and no shared structure.

Julian Esposito, "Taking inspiration from 'Battle Rap' into Copywriting", LinkedIn. Could not verify. LinkedIn blocks automated fetching, so the date and contents are unchecked. Reported as found, not as evidence. That's the rule working.

Wider than that, hip hop applied to marketing is an established genre with real names in it, including Marcus Collins in Forbes, August 2023.

The name. "Battle-tested" is a common English idiom in wide general use. Nobody owns it, which means no collision and no claim.

The structure. Six rounds. Scout the room, find the angle, build the scheme, write the bars, the rebuttal layer, crowd test and cut. Plus the rule that anything scoring low on truth dies regardless of how good it sounds. Nothing matching this was found anywhere.

WHAT YOU OWE

Nothing. Convergence, on the visible record.

But know Patton and Esposito exist, because eventually somebody in a comment section is going to bring one of them up, and having read them first is a five-minute investment.

WHAT IS ACTUALLY YOURS

The six-round structure. The truth kill rule. The crowd test as a scoring mechanism rather than a metaphor. And the decision to make the skill fire on "this headline is forgettable" instead of on the word "battle rap," which is a design choice, not an idea, and is the reason it works.

COVERAGE

Searched: the indexed web, publications, Medium, Forbes. Could not search: books, podcasts, conference talks, paywalled material, private communities, anything verbal. Could not verify one direct hit due to robots.txt. Priority also can't be fully established because my own dated post on this isn't machine-readable.

A clean result means the visible portion is clean.

That last paragraph is the point of the whole skill. The satisfying output would have been "you're original." The honest one lists two people who got to the neighborhood first, one source it couldn't open, and the specific reason the conclusion is provisional.

What's in the repo

SKILL.md                       the workflow, verdicts, the two rules, scope limits
references/verdicts.md         sampling vs converging vs biting, and how to tell
references/exposure.md         reconstructing where an idea came from, cryptomnesia
references/decompose.md        splitting a claim into components with separate lineage
references/searching.md        where to look by domain, and what cannot be searched
references/credit.md           how to credit, where to put it, and over-crediting
assets/clearance-sheet.md      the worksheet
assets/credit-lines.md         fill-in credit formats

Not legal advice

Copyright, trademark, and patent are separate questions with real consequences. When the answer might turn on one of them, this says so and points at a lawyer.

Related

battle-tested writes the line. versus picks the fight. subtext reads what came back. sequencing puts it in order. sampled checks whose it was.

License

MIT. Use it, fork it, ship it.

About

Don't bite someone else's style. A Claude Skill that checks whether the idea you're about to publish already belongs to someone, and what you owe them. Sources, dates, and an honest coverage statement every time.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors