no-slop
Skill1 starsChanged 4 months agoMITView on GitHub
Anti-slop enforcement for ANY user-facing output — landing pages, app screens, dashboards, emails, marketing. Loads before writing the first line of UI markup or copy. Triggers - building/redesigning any UI, "make a landing page", "design a screen", reviewing a UI for genericness, or /no-slop audit on an existing page.
14 other projects use a version of this.
Adopted by
Most recently taken up by
- memcpython/no-sloptheir own version, first seen Nov 15, 2026
- Tariux/AI-Skills-Not-Awesometheir own version, first seen Nov 15, 2026
- aayushchugh/skillstheir own version, first seen Nov 1, 2026
- RealAlexandreAI/public-skillstheir own version, first seen Oct 31, 2026
- iam25th1/no-sloptheir own version, first seen Oct 20, 2026
- razakiau/skillstheir own version, first seen Oct 20, 2026
Other versions14 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 14
- Tariux/AI-Skills-Not-Awesome★ 48same name
- saschb2b/skills★ 22same name
- memcpython/no-slop★ 16same name
- Eid0lon/zero-slop★ 13same name
- lcsxu/zero-slop★ 13same name
- Byk3y/no-slop★ 10same name
- aayushchugh/skills★ 4same name
- razakiau/skills★ 3same name
- eeshsaxena/agent-skills★ 2same name
- saschb2b/okf-bundles★ 2same name
4 more not listed
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 repository9 other files it gives its agents.
- skills/nirmaan/CLAUDE.mdCLAUDE.md
- brainSkill
- fable-brainSkill
- full-pictureSkill
- graphifySkill
- kimi-webbridgeSkill
- neuronSkill
- nirmaanSkill
- researchSkill
The file
We split this skill into the rules it gives an agent
4 rules. 0 are common, 3 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 14 edited versions we read keep these.
3
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 11
BEFORE generating. This skill is those constraints, as procedure.
- Distinctive: line 26
No real content available → write it now with the owner, then design.
- Very specific: line 51
All four states per view (loading/empty/error/success) · microcopy names the object ("Generate payment link", never "Submit"/"Get Started") …
- Distinctive: line 59
If yes, return to gates 2–4.
“Common” is measured: kept by at least 6 in 10 of the 14 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
/no-slop — Nothing Generic Leaves This System
Slop is the median of AI training data: what a model produces when nothing constrains it. It is not fixed by "try harder" — only by constraints applied BEFORE generating. This skill is those constraints, as procedure. Protocol authority: vault W13 (UI anti-slop) + W14 (UX) + C09 (taste); this skill is their executable form.
Reference packs (read the ones the task needs, not all)
references/BANNED.md— the slop blacklist + the tells. Read for ANY build.references/GOOD-UI.md— genre archetypes, token systems, what "good" is made of.references/UX-TRUST.md— behavior, friction, and trust/reliability patterns.
The procedure (six gates, in order — skipping one reintroduces slop)
1. Content before pixels
Obtain or write the REAL content first: actual product names, prices, field labels, realistic data lengths, the actual value proposition in the owner's words. Lorem ipsum and placeholder copy are the root cause of generic layouts. No real content available → write it now with the owner, then design.
2. Name the genre + density (out loud, one line)
operator tool (dense, tabular, keyboard-fast) · consumer commerce (trust, imagery, few choices) · field PWA (huge targets, sunlight, one hand) · SaaS landing (one job: convert one audience) · content/editorial. The genre decides density, and density is the #1 slop tell — cards-and- whitespace applied to an admin panel is slop even with perfect colors.
3. Pick ONE real reference product (same genre) and take its DNA
Name it. Then extract its actual design DNA — density, nav pattern, type scale, restraint. Token-cheap extraction: use the kimi-webbridge skill (real-browser daemon, returns text/CSS) to pull the reference's computed styles — font stacks, spacing rhythm, color values — instead of screenshotting with the in-app browser (token-expensive). One SPECIFIC reference prevents regression to the median of all references; from its DNA copy structure and discipline, never the brand look.
4. Fix tokens BEFORE components (write them at the top of the work)
ONE accent (from the owner's brand — never default purple/indigo), one neutral scale, one deliberate font pairing (state why), spacing scale 4/8/12/16/24/32/48, ONE corner radius, max 2 shadow levels, motion 150–250ms ease-out on user actions only. Every component uses only these. A token not written down will be improvised — and improvisation regresses to slop.
5. Build with the litany
All four states per view (loading/empty/error/success) · microcopy names the object ("Generate payment link", never "Submit"/"Get Started") · real content from gate 1 in every mock · zero decoration without function · check BANNED.md line by line against what you just wrote.
6. The acceptance tests (all three, mandatory)
- Squint test: blur — hierarchy must survive.
- Anonymity test: screenshot — "could this belong to any random startup?" If yes, return to gates 2–4; a good UI is recognizably THIS product's.
- Cold-user test (UX): one person completes the primary task with zero help; every hesitation is a defect (W14).
/no-slop audit mode (existing page)
Load BANNED.md → walk the page against it → report violations ranked by visibility, each with the specific fix (token change, copy rewrite, density correction). Then run the three acceptance tests. Output: a fix list, not an essay.
Copy/text slop (applies to ALL user-facing text, not just UI)
Banned in product copy and marketing: "Empower", "Seamless", "Unlock", "Elevate", "Revolutionize", "Game-changer", emoji bullets in professional surfaces (🚀✨💡), exclamation stacking, fake urgency ("Only 3 left!" when false), and benefit claims with no number or mechanism. Every headline must pass: "does this say something a competitor could NOT paste onto their site?"
Failure this prevents
The purple-gradient dark-theme emoji landing page; admin panels dressed as marketing sites; UIs that scream "AI-generated" and burn user trust on sight.
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.

