agentleFS
Fadi88Fadi88/one-man-startup/.cursor/rules/skill-human-design.mdc

one-man-startup / rules

Cursor rule16 starsChanged 2 days agoMITView on GitHub

Keeps every screen, icon, image and line of interface copy from looking AI-generated - the tells to refuse, what to do instead, and a render-and-inspect gate that the reviewer repeats blind. Use with frontend-design on every design and every ticket that changes what users see.

We split this skill-human-design.mdc into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    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
  1. Very specific:

    Before drawing, write two lines in ux.md (or ux-<platform>.md): who is using it and in what moment, and what this screen must make easy for …

    line 34
  2. Distinctive:

    Look at each in light and dark, in every language and reading direction the product has, and at the largest text size.

    line 43

≈ “Common” is an estimate: at least 1 in 5 of the 6,350 Cursor rule files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
skill-human-design.mdc
---
description: "Keeps every screen, icon, image and line of interface copy from looking AI-generated - the tells to refuse, what to do instead, and a render-and-inspect gate that the reviewer repeats blind. Use with frontend-design on every design and every ticket that changes what users see."
alwaysApply: false
---

<!-- Generated by tools/sync_adapters.py from skills/human-design/SKILL.md. Edit the source, not this file. -->

# Human design

The design humanizer: `frontend-design` sets the direction, this skill refuses the defaults that
make a product look machine-made. A design is not done until it passes the gate at the end.

First written for a real project made from this template, from two public rule lists for the
same job: anti-ai-slop (https://github.com/Vinayak-Shukla-03/anti-ai-slop) and Hallmark
(https://github.com/Nutlope/hallmark), both MIT-licensed (GitHub, checked 2026-10-17). The
wording here is our own.

## The tells, and what to do instead

| Refuse | Instead |
|---|---|
| The toolkit's default colours left as they are, and gradients by reflex | A palette from `docs/design/brand.md`, one accent with a job |
| One typeface at one weight everywhere; a second script (Arabic, say) only as a fallback | A type scale with real contrast in size and weight; faces chosen to sit together in every language the product uses |
| Every block the same rounded card with the same padding | Size, spacing and radius that show what matters most; a list where a list is clearer |
| Emoji as icons, mixed icon sets, sparkle or magic icons | One icon set, its licence recorded, one stroke and size system |
| Filler copy ("unlock", "elevate", "seamless", "your journey", "not X, it's Y"), hype, exclamation marks | Plain words the audience would use themselves |
| Placeholder data: round numbers, lorem ipsum, made-up stats or testimonials | Realistic sample data with its rough edges; any real number has a source |
| Generic 3D blobs, abstract waves, AI-generated people, mascot empty states | No image unless it earns its place; real photography or drawn illustration with its licence; empty states that say what to do next |
| Fade-in on everything, motion for decoration | Motion only where it explains a change, and off under reduced motion |
| The same screen shape as every template in the category | Structure from the task and the moment the user is in |

## Made for these people

Before drawing, write two lines in `ux.md` (or `ux-<platform>.md`): who is using it and in what
moment, and what this screen must make easy for them. After drawing, compare it with three real products in the
category (from `research.md`) and name one thing that is ours and not theirs. If nothing is,
redesign.

## The gate

1. Render the screens: the mockups with `python tools/render_mockups.py`, then the build on a
   device, emulator or browser.
2. Look at each in light and dark, in every language and reading direction the product has,
   and at the largest text size.
3. Walk the table above screen by screen, and write a pass or the tell found in
   `docs/work/<id>/human-design.md`, with the screenshot paths.
4. `reviewer` repeats step 3 without reading yours first, then compares. Any tell either of you
   finds blocks the hand-off until it is fixed.
5. The founder sees the screenshots before a design goes to build, shown to them as
   `skills/founder-voice/` says, never as paths.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.