Distinct

Skill elonmust26/distinct

An Agent Skill that fights AI homogenization — forces your AI assistant to check what exists and produce distinctive names and designs instead of the same generic defaults it gives everyone.

Install

npx -y skills add elonmust26/distinct

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Use this skill whenever the user is naming, designing, or conceiving a new product, app, extension, brand, website, or project — anytime originality matters and a generic, lookalike result would be a problem. Trigger it for "name my app", "design a landing page", "come up with a brand", "build an extension/SaaS/tool", "what should I call this", or any creative build where the user could end up with something that resembles ten other things. AI assistants tend to converge on the same default names (calm+X, X+OS, X+AI, smart+X), the same dark-gradient + sage/indigo palettes, and the same layouts — this skill actively fights that homogenization and pushes for a distinctive result. Use it even if the user doesn't say the word "original".

SKILL.md

5.5 KB, as published. Nobody here has run it

distinct — fight AI homogenization

Why this skill exists

When many different people ask an AI the same kind of question ("name my finance app", "design a clean landing page"), they tend to get similar answers — because the model converges on the same statistical defaults. The result: products that look and sound like ten others, sometimes launched within days of each other. This is not a privacy leak; it's homogenization, a known weakness of AI output ("AI slop"). The cost is real — a generic name collides with existing products, and a default design is instantly forgettable.

This skill's job: whenever the user is creating something, refuse the obvious default and steer hard toward a distinctive result. Originality is the deliverable, not a bonus.

The core rule

Before proposing any name, concept, or visual design, do the three checks below, in order. Never skip straight to a suggestion.

1. Check what already exists (mandatory, first)

If web search is available, search before naming or designing. Look up the proposed name and close variants (.com, .app, .io, .cloud, hyphenated, spaced), the product category ("X extension", "X tool"), and the top 3–5 competitors.
Report briefly what exists. If the idea or name is already crowded, say so plainly and pivot.
If search is unavailable, state that you can't verify, and lean even harder on distinctiveness (steps 2–3) to reduce collision odds.

2. Reject generic names — generate distinctive ones

Banned by default (these are the patterns AI overuses; flag and avoid unless the user insists):

calm/smart/easy/quick/super/pro + a common noun (calmfeed, smarttab, easynote)
noun + OS / AI / ify / ly / hub / flow / genius (financeos, writely, taskhub, dataflow)
literal descriptions ("AI Resume Builder", "Tab Manager Pro")
two stacked dictionary words that everyone reaches for (TaskMaster, NoteWise)

Generate distinctive candidates instead, using techniques that produce names others won't independently land on:

real but unexpected words from an adjacent field (objects, mythology, geography, crafts, nature specifics — not "leaf", but "sorrel", "kestrel", "tinder")
invented but pronounceable coinages (not random — rooted in a sound or meaning)
a concrete metaphor specific to the product's job, not its category
short, ownable, checkable as a domain

Always propose at least 5 candidates spanning different directions, and for each, note the angle and (if searched) whether it looks taken.

3. Break out of default design

When designing UI/landing/brand, do not auto-reach for the house defaults. The overused AI look to avoid unless the user wants it:

dark background + neon/gradient glow
sage-green / indigo / violet accent on near-black
the same "Inter/Fraunces, big hero, centered tagline, three feature cards" layout
generic leaf / abstract-blob / gradient-orb logos

Instead:

Ask or infer a deliberate visual direction (era, mood, an unexpected reference — editorial print, brutalist, 70s warmth, terminal, hand-drawn, Swiss, risograph...).
Offer 2–3 genuinely different directions, not one default with color swaps.
Tie the choice to the product's character, and name the references so the user can react.
Pick distinctive specifics: an unusual but legible type pairing, a non-default palette, a layout that isn't the standard hero-stack.

Output pattern

When the user asks you to name/design/conceive something, respond in this shape:

What exists — quick scan of competitors / taken names (or a note that you couldn't verify).
Distinctive options — 5+ names across different directions, or 2–3 distinct design directions, each with its angle.
How this differs — one line stating explicitly what makes the chosen direction unlike the existing field.
Then proceed with the build.

Self-check before delivering

Ask yourself, honestly:

Would three other people asking a similar question plausibly get this same name/design? → If yes, it's too generic. Redo.
Did I reach for a banned pattern (calm+X, X+OS, dark+gradient, leaf logo)? → Replace it.
Did I check what's already out there, or at least flag that I couldn't? → If not, do it now.
Can I state in one sentence what makes this distinct? → If not, it isn't.

Honest scope

This skill reduces lookalike results — it does not guarantee uniqueness. The same real-world need can lead different people to similar ideas independently, and that's outside any tool's control. What this skill controls is the part that's controllable: not defaulting to the obvious, checking the landscape, and choosing distinctive names and designs on purpose.