agent-website-design-skills
AGENTS.md2 starsChanged 4 months agoMITView on GitHub
This repository contains reusable website design skills for coding agents. Keep this repo easy for humans and agents to inspect, install, validate, and trust. The primary artifact is each skill's SKILL.md; repo-level files support discovery, distribution, and contribution quality.
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 32,671 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository10 other files it gives its agents.
- llms.txtllms.txt
- agentic-site-buildingSkill
- conversion-accessibility-polishSkill
- landing-page-craftSkill
- production-web-architectureSkill
- product-motion-designSkill
- responsive-ui-qaSkill
- technical-seo-discoverySkill
- web-design-systemSkill
- web-visual-directionSkill
The file
Show its rules and text
We split this AGENTS.md into the rules it gives an agent
18 rules. About 0 are common, 5 are distinctive, and 13 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
Distinctive · estimate
This author's own choices. Worth a look.
13
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 7
Keep this repo easy for humans and agents to inspect, install, validate, and trust.
- Very specific: line 12
Read that skill's SKILL.md.
- Very specific: line 13
Read files under references/ only when the skill points to them.
- Distinctive: line 15
Follow the selected skill's delivery contract.
- Very specific: line 20
Use web-visual-direction before implementation when the site lacks a clear design concept.
- Very specific: line 21
Use web-design-system when creating tokens, components, or consistency rules.
- Very specific: line 22
Use landing-page-craft for marketing pages, homepages, product pages, waitlists, pricing pages, and launch pages.
- Very specific: line 23
Use responsive-ui-qa after frontend changes or when layout bugs are reported.
- Very specific: line 24
Use conversion-accessibility-polish near the end of a build to improve clarity, trust, forms, CTAs, and accessibility.
- Very specific: line 25
Use agentic-site-building for natural-language code editing, orchestration, verifier loops, generated-code previews, and rollback.
- Very specific: line 26
Use product-motion-design for purposeful GSAP, scroll, micro-interaction, or Three.js work.
- Very specific: line 27
Use production-web-architecture for scalability, reliability, security, queues, caches, observability, and cost decisions.
- Very specific: line 28
Use technical-seo-discovery for crawlability, indexing, redirects, canonicals, sitemaps, rendering, and search-console incidents.
- Distinctive: line 32
Keep skill bodies concise.
- Distinctive: line 33
Do not add per-skill README files.
- Very specific: line 36
Keep skills.json and llms.txt in sync when adding, renaming, or removing skills.
- Distinctive: line 37
Do not add executable files inside skill folders.
- Distinctive: line 38
Keep repository automation deterministic, dependency-light, reviewed, and documented.
≈ “Common” is an estimate: at least 1 in 5 of the 41,581 AGENTS.md 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.
Files
Agent Instructions
This repository contains reusable website design skills for coding agents.
Repository Goal
Keep this repo easy for humans and agents to inspect, install, validate, and trust. The primary artifact is each skill's SKILL.md; repo-level files support discovery, distribution, and contribution quality.
How To Use
- Choose the smallest skill that matches the user request.
- Read that skill's
SKILL.md. - Read files under
references/only when the skill points to them. - Apply the skill to the user's target repository or page.
- Follow the selected skill's delivery contract.
- Validate frontend changes with browser evidence at relevant routes, states, and viewports when possible.
Skill Selection
- Use
web-visual-directionbefore implementation when the site lacks a clear design concept. - Use
web-design-systemwhen creating tokens, components, or consistency rules. - Use
landing-page-craftfor marketing pages, homepages, product pages, waitlists, pricing pages, and launch pages. - Use
responsive-ui-qaafter frontend changes or when layout bugs are reported. - Use
conversion-accessibility-polishnear the end of a build to improve clarity, trust, forms, CTAs, and accessibility. - Use
agentic-site-buildingfor natural-language code editing, orchestration, verifier loops, generated-code previews, and rollback. - Use
product-motion-designfor purposeful GSAP, scroll, micro-interaction, or Three.js work. - Use
production-web-architecturefor scalability, reliability, security, queues, caches, observability, and cost decisions. - Use
technical-seo-discoveryfor crawlability, indexing, redirects, canonicals, sitemaps, rendering, and search-console incidents.
Constraints
- Keep skill bodies concise.
- Do not add per-skill README files.
- Preserve each skill's frontmatter fields:
nameanddescription. - Validate changed skills with the Codex skill validator when available.
- Keep
skills.jsonandllms.txtin sync when adding, renaming, or removing skills. - Do not add executable files inside skill folders.
- Keep repository automation deterministic, dependency-light, reviewed, and documented.
Validation Commands
python scripts/validate_repo.py
npx skills add . --list
python ~/.codex/skills/.system/skill-creator/scripts/quick_validate.py ./landing-page-craftDoes 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.

