Build a website you are proud of, even if you have never written a line of code.
v4.0.0 · 4 skills · 67 commands · 135 reference docs · 15 audit sub-agents
NullToHero is an add-on for Claude. Install it once, then ask Claude in plain language to design your pages, get them ranking on Google, judge the whole site before you publish, and hand the finished thing to the person who owns it. Claude does the expert work, you stay in control.
What it is · See it built · Pick your goal · Install · Skills · Compare · Workflow · Assets
Claude already writes code. NullToHero gives it the taste and the checklists of a senior web team: a designer, an SEO specialist, a reviewer who judges the whole site at once, and the person who hands the keys to its owner.
You do not learn commands by heart. You say what you want ("make this landing page look more premium", "why am I not on Google", "is this ready to ship"), and Claude picks the right tool. The sections below show what each tool produces so you know what to expect.
nulltohero.netlify.app is the plugin's own site, and it is the honest answer to "what can I actually get out of this". Nothing on it is a mockup: it is a scroll-driven story that writes the logo onto a blank sheet, opens a terminal, raises a generic template page, then audits that page on screen and corrects it.
Open the live version: the motion is scroll-driven, so it answers your wheel rather than a timer, and the capture above only walks it at one fixed pace
The site holds itself to the rules it ships. Its source carries the plugin's own audit
annotations: where a low-contrast element is deliberate, it declares
data-contrast-exempt-reason and says why, which is the same mechanism /audit reads
when it decides whether a contrast finding is a defect or a documented exception. The one
place it shows unreadable text on purpose is the specimen page being audited, and that is
written down in the code rather than left for a reviewer to guess.
You can check the claim rather than take it:
node null-to-hero/tools/audit/gate.mjs https://nulltohero.netlify.app/ --min-score 90
# deterministic score: 100/100 FAIL: 0 WARN: 0 critical FAIL: 0
# RESULT: PASS| Step | Commands | What came out |
|---|---|---|
| 1. Frame the brief | /siteasy research |
The audience, what they need to understand in the first ten seconds, and the references worth beating |
| 2. Set the direction | /siteasy concept |
PRODUCT.md and DIRECTION.md: the idea, the anti-reference, the one signature moment, the single accent colour |
| 3. Shape before code | /siteasy shape |
The section order and the seven acts, validated as a brief before a line was written |
| 4. Build the acts | /siteasy build, /siteasy parallax |
The pinned scroll track, the act transitions, the reduced-motion fallback |
| 5. Set the type and colour | /siteasy typeset, /siteasy layout |
The type scale, the spacing rhythm, the one-accent discipline held across every act |
| 6. Look at it | /siteasy preview |
Desktop and phone screenshots at real viewports, read back and fixed in a loop |
| 7. Gate it | /audit full, /siteasy fix |
Findings routed to the command that fixes each one, then re-run until the gate came back clean |
The brief that started it was one sentence: show what the plugin does best, Awwwards-grade, with parallax and scrollytelling. Steps 1 to 3 produced no visible output at all, which is the point of doing them first.
| I want to | Type this | What you get back |
|---|---|---|
| Start from nothing | /siteasy express "a coffee shop site" |
Brief to styled landing page: concept, tokens, build, checks |
| Build a page or component | /siteasy build |
Real, production-ready front-end that matches your brand file |
| Make it better (bland, busy, static, off) | /siteasy improve index.html |
The right axis picked from your complaint and applied |
| Check the whole site | /audit yoursite.com |
One site health score and one merged action plan |
| Fix what the audit found | /siteasy fix |
Findings executed batch by batch through the remediation map |
| Rework an existing site | /siteasy overhaul yoursite.com |
Baseline, fixes, before/after proof the score moved |
| Finish and ship | /siteasy ship |
Polish, defect scan, deterministic audit and hardening, in order |
| Be found on Google and in AI answers | /seo yoursite.com |
A scored report and a prioritized action plan |
| Get a client-ready report | /audit report |
Deliverable Markdown, self-contained HTML page, or PDF |
| See it the way a real browser does | /siteasy preview index.html |
Desktop and mobile screenshots, bugs fixed in a loop |
| Hand the site to its owner | /cms entrust |
Their words become editable fields, the code stays out of reach |
NullToHero is a Claude Code plugin and a marketplace in one repository. The marketplace manifest sits at the repository root; the plugin itself lives in the null-to-hero/ folder.
A. From the marketplace (recommended, auto-updates)
/plugin marketplace add MariusYvard/NullToHero
/plugin install null-to-hero@null-to-hero-marketplace
The four skills also run on OpenAI Codex and on Kimi Code. Both read the
Agent Skills format, so the same source
serves all three hosts; null-to-hero/tools/build-dist.mjs generates each
package into dist/.
git clone https://github.com/MariusYvard/NullToHero.git
cd NullToHero
bash install.sh --target codex # or --target kimi, or --target all
On Windows: powershell -ExecutionPolicy Bypass -File install.ps1 -Target codex.
Codex gets ~/.agents/skills and ~/.codex/agents. Kimi Code gets
~/.kimi-code/skills and ~/.kimi-code/agents, and finds its fifteen
sub-agents without any launch flag. The skills install as nth-seo,
nth-siteasy, nth-audit and nth-cms, because a skills directory is
shared with every other pack on the machine and audit is a name someone else
will claim. The deterministic tools and the asset library are read from the
clone, so keep it where it is.
Claude Code is unaffected by any of this. It keeps loading null-to-hero/
directly, the prose keeps naming Claude's tools, and the build substitutes them
only in the generated packages. tests/portability.mjs fails if that stops
being true.
Both hosts were installed and run against these packages on 2026-08-19, Codex
0.148.0 and Kimi Code 0.37.2. dist/VERIFY.md records what that run showed;
bash tests/verify-hosts.sh reproduces it in about three minutes.
Updating never requires uninstalling. Refresh the marketplace, then update the plugin:
/plugin marketplace update null-to-hero-marketplace
/plugin update null-to-hero@null-to-hero-marketplace
Run /reload-plugins, or start a new session, for the new version to load.
Manual install (macOS, Linux, Windows)
B. Manual install (macOS, Linux)
git clone https://github.com/MariusYvard/NullToHero.git
bash NullToHero/install.shC. Manual install (Windows PowerShell)
git clone https://github.com/MariusYvard/NullToHero.git
powershell -ExecutionPolicy Bypass -File NullToHero/install.ps1[!WARNING] A one-liner (
bash <(curl -fsSL https://raw.githubusercontent.com/MariusYvard/NullToHero/main/install.sh)) also works, but it runs a remote script directly. Clone and readinstall.shfirst if you want to inspect it.
Tip
The short forms /siteasy, /seo, /audit and /cms work as long as no other plugin claims the same name. If you run several plugins, use the namespaced form /null-to-hero:siteasy.
The doors above are the way in. The skills below are the full reference behind them: every command stays callable on its own.
|
|
|
|
|
|
Your design partner. It plans the look, builds the pages, fixes spacing and type, makes everything responsive, and adds tasteful motion. You describe the goal, it produces real, production-ready front-end.
All 34 commands
| Command | What it does |
|---|---|
build [feature] |
Shape, then build a feature end-to-end |
shape [feature] |
Shape the UX/UI before writing code |
setup |
Create PRODUCT.md and DESIGN.md context |
concept [project] |
Set the creative direction before building: idea, anti-reference, signature moment |
research [scope] |
UX research planning, methods selection, persona and journey synthesis; generates empathy maps, journey maps and service blueprints |
ia [target] |
Information architecture, card sorting, tree testing, navigation patterns |
document |
Generate DESIGN.md from existing project code |
extract [target] |
Pull reusable tokens and components into a design system; the handoff deliverable emits the developer spec (layout, tokens, props, states, breakpoints, motion, accessibility) |
tokens [project] |
Audit or create a two-layer CSS token system, primitives + semantic layer + dark mode |
critique [target] |
UX design review with heuristic scoring |
audit [target] |
Technical quality checks (a11y, perf, responsive, WCAG 2.2, image strategy, forms) |
improve [target] |
One door for "make it better": symptom-to-axis dispatch to the right refine or enhance pass |
fix [target] |
Execute audit findings by remediation route: triage, per-command batches, verify |
polish [target] |
Final quality pass before shipping |
amplify [target] |
Amplify safe or bland designs, bolder typography, stronger color, more presence |
simplify [target] |
Reduce visual noise, tone down, strip to essence |
clarify [target] |
UX copy, error messages, button labels, empty states |
harden [target] |
Production hardening + performance, errors, i18n, edge cases, Core Web Vitals |
onboard [target] |
First-run flows, empty states, feature discovery, activation |
animate [target] |
Add purposeful animations and motion |
typeset [target] |
Typography audit, font selection, hierarchy |
layout [target] |
Spacing systems, visual rhythm, grid tools |
charts [target] |
Accessible data visualization: chart-type choice, a11y grades, non-color fallbacks |
adapt [target] |
Mobile/tablet/desktop/print adaptation |
mobile [target] |
Phone-specific ergonomics, thumb zone, touch targets, mobile navigation, virtual keyboards, mobile audit |
delight [target] |
Micro-interactions, personality in copy, satisfying feedback |
overdrive [target] |
View Transitions API, WebGL, scroll-driven animations |
video [target] |
Guaranteed-play decorative video: classify, transcode to a canvas-decodable asset (WASM decoder), emit the drop-in component |
parallax [target] |
Multi-layer depth, scrollytelling, AI-adaptive motion governance, WCAG 2.2.2 compliance |
live [target] |
Interactive variant mode (requires running dev server) |
ship [scope] |
Finish-and-ship pipeline: polish, defect scan, deterministic audit, hardening, final audit |
overhaul [url] |
Audit-driven rework: baseline, fix by remediation route, before/after compare, ship |
express [brief] |
Zero-to-landing: setup, concept, tokens, plan, build, motion, checks, harden |
Common runs: a new page (setup → shape → build → layout → adapt → amplify → harden), a refresh (improve → polish), after an audit (fix → ship), a design system (document → extract → tokens).
Your search expert. It audits a whole site or a single page, writes the structured data Google wants, builds sitemaps, and checks how visible you are in AI answers.
All 20 commands
| Command | What it does |
|---|---|
audit [url] |
Full site SEO audit, crawls up to 500 pages, scores 7 dimensions, outputs ACTION-PLAN.md |
page [url] |
Deep single-page analysis, title, meta, H1-H6, schema, images, content quality, search-experience alignment (intent, page-type match), score |
plan [business-type] |
Complete SEO strategy, architecture, content pillars, keyword plan, 4-phase roadmap |
technical [url] |
Technical audit, robots.txt, sitemaps, Core Web Vitals, mobile, security, JS rendering |
schema [url] |
Detect, validate, and generate Schema.org JSON-LD, Organization, Article, Product, etc. |
content [url] |
E-E-A-T analysis, readability, keyword density, AI citation readiness |
geo [url] |
AI search optimization, Google AI Overviews, ChatGPT, Perplexity, llms.txt, brand signals |
sitemap [url] |
XML sitemap validation and generation with industry-specific templates |
indexnow [url] |
Instant-indexing pings to the IndexNow participants, key setup, single/batch/sitemap submission; the fast lane into the indexes that feed AI answers |
images [url] |
Image SEO audit, alt text, formats (WebP/AVIF), lazy loading, CLS, LCP |
local [business] |
Local SEO, Google Business Profile, NAP consistency, citations, reviews, LocalBusiness schema |
hreflang [url] |
Hreflang validation and generation for multilingual and multi-region sites |
programmatic [url] |
Programmatic SEO, URL patterns, four graduated quality gates from a warning to a refusal, deduplication |
competitor-pages [url] |
"X vs Y" and "alternatives to X" pages with feature matrices, FAQ schema, conversion hooks |
cluster [keyword] |
Semantic keyword clustering, intent-based grouping, content architecture, gap analysis |
drift [url] |
SEO drift monitoring, baseline capture, change detection, history tracking |
backlinks [url] |
Backlink profile analysis via free data sources (Moz, Bing, Common Crawl, GSC) |
ecommerce [url] |
E-commerce SEO, product pages, category pages, faceted navigation, Product schema |
Common runs: new site (plan → build → technical → schema → sitemap → audit → /audit report), existing site (audit → technical → content → geo → backlinks), a page that will not rank (page → content → schema), local business (local → schema → geo), before a redesign (drift baseline → redesign → drift compare).
Runs every specialist at once across search, defects and design, then merges everything into one score and one action plan ordered by priority. The orchestration is documented in docs/ARCHITECTURE.md.
All 7 commands
| Command | What it does |
|---|---|
full [url] [scope] |
All 15 sub-agents across SEO, defects, and design; unified report + action plan. Optional scope runs one group: seo (5 SEO sub-agents), defects (4 inspect), design (6 siteasy), quick (one per group for a fast triage) |
checks [url] |
Deterministic pre-pass only: the computed checks and the 48 rules of the rules engine, plus SITE-AUDIT.json, no sub-agents |
verify [url] |
Consensus re-check: re-runs the gating dimensions (a11y, interaction, technical) K times and reconciles them by majority vote |
compare [A] [B] |
Diff two targets (before/after a site, or A vs B): per-check verdict changes and score deltas |
learnings [file] |
Review LEARNINGS.md candidates accumulated by real audits and turn accepted ones into rules, gates, laws or fixtures |
report [file] |
Format an existing audit into a client-ready report, a self-contained HTML page, or PDF |
review [target] |
Design engineering code review of a file or a paste: motion crimes, accessibility violations, forbidden patterns, a Before/After table with a score |
The deterministic pre-pass behind checks fetches the page once (optionally rendering a client-rendered SPA with Playwright), computes the objectively decidable verdicts (50 checks and the 48 rules of the rules engine: contrast, image dimensions, viewport, robots.txt, headings, titles, security headers, video hygiene, motion guards, media weight, AI crawler access, plus the declared-value laws for tap target size and spacing, base body size, line measure, feedback duration, decorative loop budget and scrub easing, plus the source-level rules for focus outlines, transition targets, z-index discipline and the rest), attaches to each one the fixWith route toward the command that fixes it, and writes a machine-readable SITE-AUDIT.json. That JSON powers a structural compare, score-over-time, and a CI gate you can drop into any repo as a GitHub Action (uses: MariusYvard/NullToHero@v3.8.0). See docs/ARCHITECTURE.md and null-to-hero/tools/audit/README.md. To analyze a live site in the browser with Claude, see docs/CLAUDE-IN-CHROME.md.
Common runs: a full pass (audit), a consensus re-check (audit verify) or a before and after diff (audit compare).
See sample output
A theme from /siteasy tokens, a drop-in :root stylesheet with WCAG-checked tokens:
:root {
--bg: #0B0B0C;
--surface: #161618;
--fg: #F5F5F4;
--accent: #6E56CF; /* on-accent 5.2:1, passes AA */
--ring: #6E56CF;
--text-lg: clamp(1.25rem, 1.1rem + 0.6vw, 1.6rem);
}Structured data from /seo schema, valid Schema.org JSON-LD ready to paste:
{
"@context": "https://schema.org",
"@type": "Organization",
"name": "Example",
"url": "https://example.com",
"logo": "https://example.com/logo.png"
}What /cms entrust does to a page, and the file the owner edits:
<h1>{{accueil.hero.titre}}</h1>
<p>{{accueil.hero.texte}}</p>{ "hero": { "titre": "Osez un nouveau regard.", "texte": "Opticien lunetier..." } }An action plan from /audit, ordered by severity:
## Action plan
### Critical
- Contrast 3.1:1 on the hero CTA. Raise the accent or darken the label.
### High
- LCP 3.8s. Preload the hero image and set fetchpriority="high".
### Medium
- Heading order skips from H2 to H4. Renumber the section.The last step of a project, and the one nobody ships. The site is finished, the client wants to change their opening hours, and the only person who can is you. /cms entrust turns the page's hardcoded prose into fields, vendors an editor at a pinned version, and puts a server-side allow-list between the browser and the repository. The owner edits words and pictures. The templates, the scripts and the stylesheets stay out of reach.
All 6 commands
| Command | What it does |
|---|---|
entrust [site] |
The whole chain, from a finished site to a repository ready to hand over |
carve [site] |
The extraction alone: propose the fields, read them back, write nothing until asked |
scaffold [site] |
Compile CONTENT.md into the editor, the bridge, the allow-list and the publish workflow |
accounts [site] |
Mint, remove and list the accounts the bridge will accept |
check [site] |
cms-lint and cms-scaffold --check in one pass, then the deployed bridge's own diagnosis |
handover [site] |
Regenerate CMS.md and read it with the person who will do the manual steps |
Common runs: a first handover (entrust), a re-read before writing (carve without --write), a new editor for the client's colleague (accounts), a check after someone edited a compiled file by hand (check).
What makes it safe to point at a client's repository is not the interface, which is a convenience, but four things underneath it. The bridge writes to a content branch and a workflow copies the allow-listed paths onto the production branch, so a compromised editor cannot reach a build script. The token is a fine-grained one without the Workflows permission, which puts .github/workflows/ beyond its reach by GitHub's rule rather than by ours being right. The editor is vendored at an exact version, so the admin page needs no external origin and nothing moves under a client's feet. And a write quota, counted from the branch's own commit history, means a loop in somebody's browser costs them a message rather than ten thousand commits.
One file is written by hand, CONTENT.md: what the owner may edit, the branch, the roles, the theme, the language. Everything else is compiled from it, and cms-lint runs 29 checks that fail when a compiled file and its source disagree. What no check can establish from a repository, the rights a token actually carries, whether the host deploys the branch you think, is listed at the end of the generated CMS.md instead of being left for someone to discover in production. Once the bridge is deployed it answers most of that itself: cms-diagnose.mjs signs in and reports, in booleans and never in values, which variables are set on the context being served, whether the token can write, whether both branches exist, and when the token expires. Three weeks before that date the bridge starts saying so in the host's log, because a token that expires stops writing on a Tuesday without warning anybody.
This chain assumes Netlify. cms-scaffold writes the bridge to netlify/functions/cms.mjs, the headers to _headers, and reads the build command from netlify.toml; the quota exists because Netlify bills build minutes. Nothing in the bridge itself is Netlify-specific, it is one request handler over the standard Request and Response, so another host that runs functions from a repository would need the three files renamed and its own way of setting environment variables. That port has not been done, and this README will not claim it has.
Measured on a real 33-page site: 712 fields extracted, and 32 of the 33 pages come back byte for byte identical after the extraction is filled back in. The one that does not holds a doubled space, and the tool names the page rather than staying quiet.
NullToHero overlaps four kinds of tool. The honest comparison: each column is excellent at what it is for, none of them spans the whole loop of building, auditing, scoring, fixing and handing over inside your own files.
Text version of the table
| Capability | NullToHero | AI site builders v0 · Lovable · Bolt |
Audit tools Lighthouse · axe |
SEO suites Semrush · Ahrefs |
Design skills ui-ux-pro-max · ux-ui-mastery · Taste |
|---|---|---|---|---|---|
| Designs and builds real pages | ✅ | ✅ | — | — | 🟡 |
| Audits an existing site | ✅ | — | ✅ | ✅ | — |
| Routes every finding to the command that fixes it | ✅ | — | — | — | — |
| Deterministic 0-100 score, reproducible in CI | ✅ | — | 🟡 | 🟡 | — |
| Design taste scored (memorability, motion, Gestalt) | ✅ | 🟡 | — | — | 🟡 |
| SEO with structured data | ✅ | — | 🟡 | ✅ | — |
| AI-search visibility (GEO, llms.txt) | ✅ | — | — | 🟡 | — |
| Accessibility: WCAG 2.2 rules and contrast checks | ✅ | 🟡 | ✅ | — | 🟡 |
| CI gate on every push | ✅ | — | ✅ | — | — |
| Runs inside Claude, on your own files | ✅ | — | — | — | ✅ |
| Hands the finished site to a non-technical owner | ✅ | — | — | — | — |
| Free, open source, no account | ✅ | — | ✅ | — | ✅ |
| Editable knowledge (rules, laws, resources as CSV) | ✅ | — | — | — | 🟡 |
✅ yes · 🟡 partial · — no. The figure above says the same thing with a tick, a half disc and a dash, so it survives greyscale. Both are drawn from null-to-hero/tools/data/compare-matrix.csv. Nuances: Lighthouse's deterministic score covers performance, not design or content; Semrush's Site Health score is deterministic but proprietary and not CI-native; builders generate tasteful UI without judging or scoring it.
NullToHero is the one that spans build, defects, SEO, a scored whole-site audit and the handover in a single plugin, with every finding routed to the command that fixes it. It is not a hosted product or a visual editor: it runs inside Claude and edits the real files in your project, so the output is yours to keep and version.
The doors are the flow. A project usually walks through six of them, in this order:
/siteasy express nothing yet: brief to a first shippable page
or
/siteasy build something exists: add the next piece
|
/siteasy improve "make it better": the right axis, one pass at a time
|
/audit the whole site, scored, every finding routed to its fix
|
/siteasy fix execute the findings, batch by batch
|
/siteasy ship polish, scans, hardening, gates: out the door
|
/cms entrust the owner can change the words, not the code
Reworking an existing site instead: /siteasy overhaul chains the baseline audit, the fixes and the before/after proof. Growing traffic after launch: /seo plan sets the strategy and names the specialist passes to run; /seo drift watches for regressions.
The same flow, specialist by specialist
/siteasy research understand the users
/siteasy ia validate the structure
/siteasy setup define brand, audience, tone
/siteasy shape shape the UX before coding
/seo plan build the SEO strategy in parallel
/seo cluster group keywords by intent
|
/siteasy build build the interface
/siteasy tokens set up the token system
/siteasy layout fix spacing and rhythm
/siteasy adapt make it responsive
/siteasy mobile tune phone ergonomics
|
/siteasy amplify make it beautiful
/siteasy simplify strip to the essence
/siteasy typeset refine the typography
/siteasy animate add motion
/siteasy delight add micro-interactions
/siteasy clarify sharpen the copy
|
/audit checks catch anti-patterns, no model involved
/siteasy preview see it in a real browser
/siteasy critique heuristic UX review
/audit review final code-quality gate
/siteasy polish last quality pass
|
/seo audit full SEO check
/seo technical crawl and render health
/seo schema add structured data
/seo content E-E-A-T and readability
/seo images image SEO
/seo geo AI-search visibility
|
/siteasy harden harden for production
/audit report client-ready report
/seo drift watch for regressions
|
/cms entrust hand the keys to the owner
Tip
In a hurry, /audit yoursite.com runs the whole check in a single pass.
Set up your project
NullToHero works best with two small files in your project root. Claude reads them so its output matches your brand.
PRODUCT.md, who your users are, your brand, tone and anti-references. Create it with/siteasy setup.DESIGN.md, your colors, typography and components. Generate it with/siteasy document.
Two more appear as you work: DIRECTION.md (the committed creative direction, written by /siteasy concept) and LOG.md (the running journal the journeys append to and resume from).
The null-to-hero/assets/ folder ships an original, license-clean starter library: 139 icons, 20 background patterns, 18 spot illustrations, 34 animations and 6 templates. Icons and patterns take the surrounding color and the animations honor prefers-reduced-motion. Everything is CC0 for the media and MIT for the templates, so it is safe to copy into any project. Open null-to-hero/assets/gallery.html to browse the whole set, and null-to-hero/assets/README.md for how to wire each kind in. During a build, /siteasy build recommends curated external sites first and uses this library as a fallback for quick, offline or placeholder assets.
NullToHero ships 135 reference docs that Claude loads only when it needs them, so a large project does not eat your context budget.
See the full knowledge base
siteasy, design (86): accessibility-engineering, adapt, animate, animation-engineering, assets-library, audit, bolder, brand, brand-identity, clarify, cognitive-load, color-and-contrast, color-systems, colorize, component-patterns, component-recipes, concept, conversion-experiments, conversion-quality, craft, creative-patterns, critique, css-architecture, dark-mode-engineering, data-viz, delight, design-tokens, distill, document, elevation, extract, fetch-asset, fix, form-patterns, gestalt, handoff, harden, heuristics-scoring, image-strategy, improve, information-architecture, inspiration, interaction-design, journey-express, journey-mapping, journey-overhaul, journey-ship, landing-patterns, layout, live, memorability, mobile-ergonomics, motion-choreography, motion-design, objections, offer-diagnostic, onboard, optimize, overdrive, parallax, personas, polish, preview, print-styles, product, quieter, resource-recipes, resource-recommendations, responsive-design, shape, ship-checklist, signature-moments, slop-patterns, sourcing-external-code, spatial-design, stock-media, style-systems, teach, testing-strategy, tokens, typeset, typography, ux-research, ux-writing, video, wcag-2-2
seo, search (27): action-plan, ai-overview-recovery, audit, backlinks, cluster, competitor-pages, content, drift, ecommerce, geo, head-meta, hreflang, images, indexnow, local, measurement, migration, page, performance, plan, privacy-consent, programmatic, schema, search-console, sitemap, sxo, technical
seo, plan assets (6): agency, ecommerce, generic, local-service, publisher, saas
audit, judging (12): checks, code-quality, compare, full, html-report, learnings, refine, rendered, report, review, rules-engine, three
cms, handing over (4): architecture, carve, entrust, operate
shared state and routing: DIRECTION.md and LOG.md project files read by every command, null-to-hero/tools/data/laws.csv (16 canonical numeric laws, CI-checked citations), null-to-hero/tools/data/remediation-map.csv routing every check and rule to the command that fixes it (fixWith in SITE-AUDIT.json), and null-to-hero/tools/reference-graph.json (the reference graph, zero orphans enforced by CI).
A stack-aware design-system generator also lives under null-to-hero/tools/design-system/, covering 16 technology stacks (React, Next.js, Vue, Svelte, Astro, Nuxt, Angular, Laravel, HTML and Tailwind, shadcn/ui, SwiftUI, React Native, Flutter, Jetpack Compose, Three.js, Nuxt UI).
Requirements
- Node.js, for
/audit checks,/siteasy preview, the whole/cmschain and the validator (tests/validate.js). - Playwright, installed on first
/siteasy previewrun, and required by/cms carve, which reads the page in a real browser and refuses rather than guessing when it is missing. - Python 3, for the design-system generator (
/siteasy setup) and the Python tests.
- Changes by version: CHANGELOG.md
- How to contribute: CONTRIBUTING.md
- Reporting a vulnerability: SECURITY.md
- Credits and third-party licenses: ATTRIBUTION.md, NOTICE
node tests/validate.js # run before opening a PRBuilt by Marius Yvard · Live showcase · Releases · Changelog · Contributing
License: Apache 2.0. See LICENSE.