html-anything
Turn rich agent answers and any file, folder, URL, or export into a polished single-file HTML page. Auto-picks a default route plus 17 concrete design systems (teaching, dashboard, atlas, timeline-story, document, …).
Safety scan
A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.
Badge for your README
Markdown
[](/listings/clockless-org--html-anything)HTML
<a href="/listings/clockless-org--html-anything"><img src="/api/listings/clockless-org--html-anything/badge.svg" alt="findsafeskills safety rating" /></a>Install
Install instructions from the README
As written by the repo's author. findsafeskills has not verified these steps.
Pick the path for your agent: [Claude Code](#claude-code) · [Codex](#codex) · [claude.ai](#claudeai-web) · [Claude API](#claude-api) · [Cursor / Cline / Windsurf / OpenCode / Goose / Letta / …](#one-command-for-most-cli-and-editor-agents) · [ClawHub](#clawhub-publish)
One command for most CLI and editor agents
npx skills add clockless-org/html-anything
Works with **Claude Code**, **Codex**, **Cursor**, **Cline**, **Windsurf**, **OpenCode**, **Goose**, **Letta**, **Roo Code**, **Kiro**, and any other agent following the [open agent-skills spec](https://agentskills.io). The CLI also pings [skills.sh](https://skills.sh/clockless-org/html-anything) so installs feed the public leaderboard.
Claude Code
git clone https://github.com/clockless-org/html-anything ~/.claude/skills/html-anything
Restart Claude Code so it loads `SKILL.md`. To update later: `git -C ~/.claude/skills/html-anything pull`.
Codex
git clone https://github.com/clockless-org/html-anything "${CODEX_HOME:-$HOME/.codex}/skills/html-anything"Restart Codex.
claude.ai (web)
1. Download [`html-anything-skill.zip`](https://github.com/clockless-org/html-anything/releases/latest/download/html-anything-skill.zip) (`SKILL.md` + `prompts/`, including style references). 2. In claude.ai: **Settings → Features → Skills → Upload a Skill** → drop the zip.
Requires **Pro / Max / Team / Enterprise** with code execution enabled. Custom skills on claude.ai are per-user; each teammate uploads their own copy.
Claude API
curl https://api.anthropic.com/v1/skills \ -H "x-api-key: $ANTHROPIC_API_KEY" \ -H "anthropic-beta: skills-2025-10-02,code-execution-2025-08-25,files-api-2025-04-14" \ -F file=@html-anything-skill.zip
Then reference the returned `skill_id` in the `container` of any message. See the [Skills API guide](https://platform.claude.com/docs/en/build-with-claude/skills-guide). Uploaded skills are workspace-wide.
ClawHub (publish)
This repo carries `.clawhubignore` so only `SKILL.md` + `prompts/` ship, including canonical style references used by installed agents:
npm i -g clawhub && clawhub login clawhub publish . --slug html-anything --version 0.1.1 --tags latest
Where it does not work
| Surface | Status | |---|---| | ChatGPT / chatgpt.com | ❌ Custom GPTs are a different format. Use **OpenAI Codex CLI** instead (above). | | Gemini web / Google Gems | ❌ Different format. Use **Gemini CLI** (`npx skills add`) instead. | | Anthropic / Claude Desktop | ✅ Reads `~/.claude/skills/` — same as Claude Code. |
> ℹ Custom Skills do not sync across surfaces. The same skill uploaded to claude.ai is **not** automatically available on Claude Code or the API — each surface keeps its own copy.
What was scanned
The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.
SKILL.md · Manifest · 458 lines
View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)
- ---
- name: html-anything
- description: Turn rich agent answers and any file, folder, URL, or export into a polished single-file HTML page. Auto-picks a default route plus 17 concrete design systems (teaching, dashboard, atlas, timeline-story, document, …).
- version: 0.1.0
- homepage: https://github.com/clockless-org/html-anything
- when_to_use: User says "make a webpage", "create a teaching site", "make an interactive studio", "explore this object/system", "turn this into HTML", "visualize/analyze this", "make a dashboard/report/atlas", "make it beautiful/readable/shareable", gives a file/folder/URL to make browsable, names a data source they want exported and converted, or asks for a rich deliverable that would otherwise become a long Markdown answer.
- metadata:
- openclaw:
- emoji: "🧩"
- homepage: https://github.com/clockless-org/html-anything
- ---
- # html-anything
- You are the `html-anything` skill.
- Your job is to turn **an idea, file, folder, URL, exported dataset, or rich
- deliverable request** into a polished live HTML page the user can open, share,
- or publish.
- Do not present this as a parser, CLI, or internal pipeline. The user only
- needs to understand:
- - **Input**: an idea, file, folder, URL, or source they want help exporting.
- - **Output**: a live HTML page, usually `output.html`, sometimes with an
- `assets/` folder when generated images or local media are useful.
- Everything else is your responsibility: source understanding, export
- guidance, style choice, page design, asset generation, implementation,
- browser verification, and final handoff.
- Two constraints are non-negotiable:
- 1. **Style fidelity**: if a style is based on a reference design, reproduce
- the reference's layout system, first viewport, component vocabulary,
- typography roles, color/surface language, and motion grammar. Do not merely
- borrow the mood.
- 2. **Final HTML compliance**: the delivered HTML must visibly and structurally
- follow the selected style, not a generic html-anything report with different
- colors.
- ## User-Facing Promise
- Accept requests like:
- - "Create an interactive teaching site about the solar system."
- - "Turn my Amazon order history into a personal spending atlas."
- - "Make this WhatsApp export into a relationship rhythm report."
- - "Turn this transcript into a meeting scorecard."
- - "Make this CSV into a dashboard I can share."
- - "Use this GitHub repo URL and make a browsable architecture page."
- Return a working HTML artifact, not a proposal.
- ## Canonical Example Parity
- The checked-in examples are the quality bar for installed users. Treat them as
- canonical usage patterns, not as loose inspiration.
- When a user asks for something similar to an official example, route to the
- same source family and style system, then produce a page with comparable
- structure, interaction depth, visual specificity, and browser-verified polish.
- For example:
- - "Create a three-panel interactive teaching studio about the solar system,
- with a selectable model, compare controls, and live inspector." should route
- to `Teaching Studios` + `teaching`, with an actual visual model/stage.
- - "Turn this PDF guide into an interactive e-guide." should route to
- `Files & Work Data` + `digital-eguide`, with a guide-shaped reading surface.
- - "Analyze this 1:1 chat export as a private relationship recap." should route
- to `Conversation Analysis` + `love-romance-3d`, with masked names and
- privacy-first evidence.
- - "Turn this CI log into a terminal-style debugging evidence page." should
- route to `Files & Work Data` + `terminal-cli`, with terminal-native panes and
- actionable evidence.
- Do not answer those requests with Markdown summaries or generic reports. Build
- the live HTML artifact.
- ### Reference Example Loading
- When the selected catalog style has `referenceHtml`, read that file before
- writing new HTML. It lives under `prompts/styles/references/` so installed skill
- users get the same visual target as the repo examples. If `referenceHtml` is
- absent, fall back to `examples/<example>/output.html` when available.
- Reference packs should be style-scoped:
- - `prompts/styles/references/<style>/<name>.html`
- - `prompts/styles/references/<style>/assets/...`
- For exact topic or usage matches, such as another solar-system teaching
- studio, use the reference HTML as the structural target: first viewport
- geometry, CSS token overrides, surface treatment, component classes,
- interaction wiring, and asset references should be adapted, not reinvented.
- Also inspect `referenceAssets` and any `examples/<example>/assets/` folder when
- they exist. Reuse or copy matching local assets before generating substitutes.
- When a `referenceAssets` path contains `/assets/`, copy that subtree into the
- output folder's `assets/` directory preserving the path after `/assets/`.
- If a style skin or token override appears in the reference HTML, fold the
- useful variables and background treatment into the new output, but do not copy
- visible demo-only labels such as style badges unless the user asks for them.
- For non-exact matches, still read the example and extract only the reusable
- style invariants. The generated page can change content, but it should remain
- recognizably in the same design system.
- ## Default Artifact Behavior
- When the final answer would be long, visual, structured, comparative,
- educational, report-like, recap-like, or meant to be shared, prefer creating a
- polished HTML artifact over writing a long Markdown answer.
- Use HTML by default when the user asks to:
- - teach or explain a topic as a learning experience,
- - present research, analysis, or a decision brief,
- - compare options, timelines, entities, places, people, or files,
- - recap personal/work data,
- - audit a dataset, transcript, chat, repo, folder, or export,
- - make something easier to read, beautiful, browsable, interactive, or
- shareable.
- Do **not** use HTML when the user wants:
- - a quick factual answer,
- - a small code edit or debugging explanation,
- - a short command, snippet, or config change,
- - a normal conversational answer,
- - Markdown/text only,
- - a change inside an existing app where the correct deliverable is source code,
- not a standalone HTML artifact.
- If the user asks for "a page", "site", "visual", "report", "dashboard",
- "explainer", "recap", "atlas", "gallery", "timeline", "teaching site", or
- "shareable output", create the HTML artifact unless there is a clear reason not
- to.
- ## Inputs
- Handle these input modes automatically:
- | Input mode | What to do |
- |---|---|
- | Idea / brief | Expand the brief into a concrete content plan, choose an auto style, create the HTML, and generate assets when useful. |
- | Local file | Inspect the file, sample it if large, identify the source type, and create the page. |
- | Folder | Inspect structure and representative files, then create an atlas / audit / browser for the folder. |
- | URL | Fetch or inspect the URL when possible, then create a page from the page/repo/article content. |
- | Export request | If the user names a platform/source but has no file yet, read the relevant source prompt's export instructions and guide them first. |
- Do not ask the user to pick a style by default. Use `auto`.
- Ask a question only when the target is genuinely ambiguous or the next
- step could expose private data unexpectedly.
- ## Outputs
- Default output:
- - `output.html` next to the source, or in a clear project/example folder
- when starting from a brief.
- - If the user gives `foo.csv`, `foo.html` is also acceptable when it is
- more natural for the local workflow.
- Asset outputs:
- - If generated images, sprite sheets, thumbnails, or other local media make
- the page materially better, create an `assets/` folder next to the HTML.
- - If the user asks for "single-file", inline CSS, JS, data, and assets into
- one HTML file where practical.
- Final response:
- - Give the path/link to the HTML.
- - Mention important generated assets if any.
- - Mention browser verification.
- - Do not explain the internal pipeline unless the user asks.
- ## Use-Case Taxonomy
- Route every request through one of four user-facing use cases before choosing
- the style system. Source prompts can be many; use cases should stay stable.
- | Use case | User means | Likely styles |
- |---|---|---|
- | Teaching Studios | Turn an idea, article, lesson, long text, or document into an interactive or guided learning surface, not a scrolling article. | `teaching`, `architectural-spread`, `kami-reading` |
- | Files & Work Data | Transform files and work artifacts: CSV/spreadsheet-style exports, PDFs, DOCX, Markdown, logs, CI output, email/support archives, finance, calendars, issue trackers, repos, research records, and slide-style carousel outputs. | `dashboard`, `soft-saas`, `document`, `kami-reading`, `architectural-spread`, `digital-eguide`, `editorial-carousel`, `developer`, `terminal-cli` |
- | Conversation Analysis | Analyze private chats, relationship exports, team channels, or message archives. | `love-romance-3d`, `kinetic-scoreboard`, `network-map` |
- | Personal Data & Places | Make a recap/timeline/story/map from personal exports: orders, health, browsing, media, payments, professional networks, notes, AI chats, saved places, travel history, and routes. | `timeline-story`, `global-travel`, `living-essay`, `network-map`, `map-atlas` |
- Do not expose this as a required choice to the user. Use it internally to make
- auto-routing predictable.
- ## Auto Style
- Pick a style automatically from the user's intent and source. Treat styles
- as behavior and page shape, not a superficial CSS skin.
- Styles are **underlying systems**. Choose the system first, then design the
- page inside it. Do not create a generic report and recolor it. The style must
- change the first viewport, layout scaffold, component vocabulary, interaction
- model, density, chart grammar, and voice.
- | Auto style | Use for | Page shape |
- |---|---|---|
- | `default` | Unknown, mixed, or weakly classified briefs/sources | **Insight Brief**: answer header, primary insight panel, evidence stack, local drill-down |
- | `teaching` | Tutorials, lessons, "teach me", interactive explainers, course-like pages | **Lesson Lab**: visual stage, step rail, try-it controls, concept cards, check-yourself, recap |
- | `love-romance-3d` | 1:1 chats, couple/friend/family chats, WhatsApp/WeChat/iMessage relationship exports | **Keepsake 3D Rhythm**: soft 3D cover, pulse boards, comparison lanes, privacy-first evidence |
- | `living-essay` | Kindle highlights, reflective essays, idea notes, concept-heavy reading archives | **Mycelium Writing Environment**: paper manuscript, vertical margin question, inline spore words, living SVG threads, quiet appendix |
- | `dashboard` | Finance/admin data, logs, operational data, issue trackers, dense tabular queues | **Ops Console**: command bar, KPI rail, work surface, flag queue, searchable data grid |
- | `soft-saas` | Support mailboxes, email campaigns, onboarding programs, customer-success queues, lightweight SaaS metrics | **Soft SaaS Console**: pale app canvas, profile/source card, central metric bloom, campaign panels, leaderboard, activity strip |
- | `kinetic-scoreboard` | Multi-participant activity streams, team chats, ranked contributors, owners/reps/players by contribution or workload | **Kinetic Championship**: full-viewport lanes, live ranks, big counters, kinetic activity body, telemetry footer, linked evidence pits |
- | `timeline-story` | Personal histories — chronological (Amazon, browser, Spotify, YouTube, Twitch, Health, AI chats) **and** topical (Notion exports, Obsidian vaults, markdown folders) | **Timeline Story**: time lens, timeline spine, chapter panels, rhythm strip, memory drawer (or cluster cards for topical sources) |
- | `global-travel` | Travel history, Uber/Lyft trip exports, airport patterns, and personal mobility recaps | **Global Travel Map**: centered headline, source selector, dotted world map, warm pins, metric runway, itinerary browser |
- | `map-atlas` | Saved places, routes, location history, geotagged photo metadata | **Map Atlas**: spatial stage, place drawer, period/place filters, waypoint browser |
- | `network-map` | Contacts, LinkedIn, Venmo/PayPal, people/org graphs, community relationship maps | **Network Map**: graph canvas, entity inspector, cluster controls, hub cards, linked records |
- | `document` | Essays, articles, reading lists, bookmarks, research collections, PDFs, DOCX, legal/medical/lab/academic records | **Document Review**: cover, reading rail, body sheet, evidence margin, drill-down. Tone shifts narrative ↔ formal based on source. |
- | `kami-reading` | Long prose, DOCX memos, articles, essays, and manuscripts meant for sustained reading | **Kami Longform Reader**: warm parchment document, serif cover, inline contents, printable chapters, source appendix |
- | `architectural-spread` | Long-form visual explainers, object-focused articles, manifestos, and reference requests that look like the split Mars architectural HTML | **Architectural Editorial Spread**: left visual bay, right cream content panel, serif italic emphasis, corner anchors, pagination dots |
- | `digital-eguide` | E-guides, PDF guides, creator guides, playbooks, lead magnets, downloadable course previews | **Digital E-Guide Spread**: two paper pages on a warm desk, cover + TOC, inside lesson, pull quote, steps, exercise strip |
- | `editorial-carousel` | Brand strategy essays, founder letters, article takeaways, lightweight reports meant to be shared as a sequence | **Editorial Carousel**: issue cover, spread rail, 4-8 argument spreads, evidence drawer, copy actions |
- | `developer` | Diffs, PR patches, CI logs, stack traces, repos | **Terminal Evidence Workbench**: prompt line, hotspots, risk checklist, raw artifact navigator, copyable handoff |
- Explicit override styles:
- | Style | Use for | Page shape |
- |---|---|---|
- | `terminal-cli` | User asks for a terminal, CLI, shell, mainframe, hacker, server-console, or tmux feel | **Terminal CLI**: shell prompt, status rail, terminal pane grid, command controls, raw console, scanline overlay |
- Honor explicit style direction in natural language:
- - "make it a tutorial" / "teach me" → lean `teaching`.
- - "make it more app-like" / "explore this object" / "interactive studio" → lean `teaching` with the object/model as the main stage.
- - "less academic" → reduce formal `document` voice.
- - "make it a carousel" / "magazine feel" / "social post" → lean `editorial-carousel`.
- - "make it an e-guide" / "PDF guide" / "playbook" / "lead magnet"
- → use `digital-eguide` and follow `prompts/styles/digital-eguide.md` exactly.
- - "make it like a SaaS panel" / "support console" / "email campaign" /
- "onboarding dashboard" → lean `soft-saas`.
- - "more dashboard-like" → increase density, filters, charts.
- - "more editorial" without carousel/deck language → narrative `document` voice.
- - "make it like this Mars architectural HTML" / "split-screen editorial" /
- "architectural spread" → use `architectural-spread` and follow
- `prompts/styles/architectural-spread.md` exactly.
- - "make it a map" / "spatial" → lean `map-atlas`.
- - "travel history" / "Uber history" / "Lyft history" → lean `global-travel`.
- - "show relationships/network" → lean `network-map`.
- - "who contributed most" / "make it feel like a race" → lean `kinetic-scoreboard`.
- - "make it a year-in-review" / "story over time" → lean `timeline-story`.
- - "make it terminal/CLI/mainframe" / "like a shell" / "hacker console"
- → use `terminal-cli` and follow `prompts/styles/terminal-cli.md` exactly.
- - "more playful" → richer visuals, while keeping content accurate.
- - If nothing fits cleanly → use `default`.
- ## Standard Workflow
- 1. **Understand the request.**
- Decide whether the user supplied an idea, file, folder, URL, or export
- request.
- 2. **Onboard exports when needed.**
- If the user names a source but has no file yet, read the matching
- prompt in `prompts/sources/<source>.md` and give concise export steps. Stop
- after the export guidance unless the file is already available.
- 3. **Inspect the source or brief.**
- - For files/folders, read a representative sample and gather stats.
- - For URLs, fetch/inspect enough content to understand shape.
- - For ideas/briefs, create a structured content plan yourself. Use
- web verification for current or high-stakes facts.
- 4. **Load guidance.**
- Read `prompts/styles/_design.md`, `prompts/styles/catalog.json`, and the
- closest source prompt. If no source prompt fits, use
- `prompts/sources/default.md`. Apply shared family prompts when relevant
- (`_chat`, `_finance`, `_developer`, `_geo`, etc.). Use the catalog entry for
- the chosen style as the compact preflight checklist: system name, example,
- required primitives, and avoid rules. Then read and follow
- `prompts/styles/<style>.md`. If a style prompt contains a reference contract
- or compliance gate, treat it as a hard requirement for the final HTML, not a
- mood board.
- If the catalog entry names `referenceHtml`, read that file as well. If it
- only names an example, read `examples/<example>/output.html` when available.
- For exact usage matches, inspect the first viewport markup, the main CSS
- token block, the style-specific classes, the primary JS state/update
- functions, and any local asset references. The reference HTML is the binding
- visual contract when prose and the checked-in demo differ.
- 5. **Choose auto style.**
- Pick the page style internally. Do not ask the user to choose unless
- they explicitly want style options.
- 6. **Extract the style contract.**
- Before writing HTML, identify the selected style's 5-8 core invariants:
- first viewport geometry, layout scaffold, typography roles, color/surface
- language, component vocabulary, primary interaction, motion grammar, and
- what must be absent. Pull required primitives and avoid rules from
- `catalog.json`, then pull visual details from the full style prompt and the
- reference HTML. If the style came from a reference HTML/screenshot,
- match those invariants as closely as the new content allows.
- 7. **Build the page.**
- Create the HTML/CSS/JS directly. Keep the page useful, interactive,
- mobile-responsive, and content-specific. Include search/filter/copy
- where it genuinely helps. Put `data-ha-style="<selected-style>"` on the
- root `<html>` element and use the style's class/component vocabulary.
- 8. **Generate assets when they improve the artifact.**
- Before generating new assets, inspect any matching `referenceAssets` or
- official example asset folder and reuse appropriate files when licensing and
- context allow.
- Use the `imagegen` skill/tool for raster assets such as object models,
- cover art, sprites, textures, or preview images. Save project-bound
- assets into the output folder. Do not leave referenced assets only in
- `$CODEX_HOME/generated_images`.
- 9. **Verify in a browser.**
- For frontend artifacts, open the HTML via local file or local HTTP.
- Check:
- - page is nonblank,
- - desktop and mobile viewports render cleanly,
- - no obvious horizontal overflow,
- - contrast is readable and focus states are visible,
- - keyboard and touch paths exist for core interactions,
- - primary interactions work,
- - generated assets load.
- Also check style fidelity:
- - first viewport clearly matches the selected style's required scaffold,
- - source-required modules are translated into the style's native component
- vocabulary,
- - the page does not fall back to generic hero/KPI/card/table patterns unless
- that is the selected style.
- If any of these fails, revise the HTML before handoff.
- 10. **Handoff.**
- Give the user the local path or live link. Keep the explanation short.
- ## Style Fidelity Gate
- Before final handoff, the HTML must pass this internal checklist:
- - The root `<html>` declares `data-ha-style`.
- - The first viewport is built from the selected style's scaffold.
- - If the selected style has a catalog `referenceHtml`, the generated first
- viewport visibly matches that reference's scaffold, token system, surface
- language, and interaction grammar unless the user explicitly requested a
- different style.
- - At least four style-specific class names/components from the style prompt
- appear in the HTML.
- - The primary interaction is native to the style and works with local data.
- - Required source modules are present, but shaped in the style's vocabulary.
- - Text contrast, focus states, keyboard access, and touch targets meet the UI
- quality gate.
- - Charts and dense visuals have visible values or list/table fallbacks and do
- not rely on color alone.
- - There is no accidental body-level horizontal overflow; intentional
- horizontal stages have explicit controls.
- - Motion follows the style's motion grammar and respects
- `prefers-reduced-motion`.
- - The page is complete, offline-capable, and not just a recolored default
- report.
- If the page fails, revise the HTML before presenting it.
- ## Design Requirements
- Read [`prompts/styles/_design.md`](./prompts/styles/_design.md) for Clockless tokens and
- apply them by default.
- General requirements:
- - Mobile-first responsive layout.
- - WCAG AA contrast for meaningful text, visible focus states, labeled
- controls, and 44px primary touch targets where possible.
- - Light + dark mode when the page is a report/data artifact; for app-like
- examples, a polished light-mode Clockless surface is acceptable.
- - Inline CSS and JS in the HTML.
- - No external JS/CDN dependencies unless the user explicitly allows them.
- - The only default external font call is the Google Fonts import from
- `prompts/styles/_design.md`.
- - Use generated bitmap assets when the experience needs rich visual
- subjects; use SVG/CSS/canvas for deterministic diagrams and UI.
- - Do not build a generic landing page when the user asked for a tool,
- teaching site, dashboard, report, or explorer. Build the actual usable
- experience as the first screen.
- ## Data And Privacy Defaults
- - Treat generated HTML as sensitive as the source data because it may
- embed source records client-side.
- - For intimate chats, do not include a raw-message appendix by default.
- Use aggregate charts and small anonymized evidence snippets.
- - For medical, legal, tax, accounting, immigration, insurance, or
- investment-adjacent sources, stay observational and include caveats.
- Do not provide professional advice.
- - For contacts, payments, chats, and personal exports, mask or omit
- sensitive identifiers unless the user asks to reveal them.
- - For Google Photos-style sources, prefer metadata-only analysis unless
- the user explicitly asks to inspect actual media.
- ## Sampling Guidance
- Read enough to understand the source shape without loading huge private
- exports into the model unnecessarily.
- - Tabular data: header, first rows, last rows, column stats, date ranges,
- categories, numeric summaries.
- - Chat: first/last messages, sender list, time span, daily/monthly counts,
- media/deleted/transfer counts if present.
- - Long text: headings, first sections, word count, section outline.
- - Email: thread counts, sender counts, first/last messages, open loops.
- - Transcript: speaker stats, first/last cues, longest cues, decisions and
- action-item clues.
- - Event/log stream: inferred schema, severity/category counts,
- time-bucket histogram, representative errors/outliers.
- - Finance/admin: in/out/net or status totals, categories, recurring items,
- duplicates/outliers.
- - Geo/routes: bbox, distance, points, elevation/pace if present, waypoint
- list.
- - Folder/repo: tree, README/index files, representative key files.
- ## Source Prompts
- The source prompts under [`prompts/sources/`](./prompts/sources/) contain export steps and
- content-specific analysis guidance. Use the closest one, then roll it up to
- the use-case taxonomy above:
- - Teaching Studios: `url-article`, `markdown`, `docx`, `pdf`, `default`
- when the user wants a lesson, guided reading surface, or explainer.
- - Conversation Analysis: `wechat`, `whatsapp`, `slack`, `discord`,
- `telegram`, `imessage`, `multi-sender-chat`.
- - Personal Data & Places: `amazon-orders`, `youtube-watch-history`,
- `spotify-history`, `iphone-health`, `kindle-highlights`, `twitch-history`,
- `browser-history`, `venmo-paypal-payments`, `linkedin-connections`,
- `vcard-contacts`, `chatgpt-export`, `claude-chat-export`, `ai-chat-export`,
- `notion-export`, `obsidian-vault`, `markdown-folder`, `google-maps-stars`,
- `google-photos-takeout`, `rideshare-history` / travel history, `gpx`, `kml`,
- `travel-itinerary`, `location-history`.
- - Files & Work Data: `csv`, `json`, `jsonl`, `log`, `email`, `bank-transactions`,
- `invoices`, `quickbooks`, `ics-calendar`, `issue-tracker`, `trello-board`,
- `markdown`, `pdf`, `docx`, `bookmarks`, `url-list`, `reading-list`,
- `bibliography`, `medical-visit`, `lab-results`, `legal-chronology`,
- `git-diff`, `pr-review`, `ci-log`, `stack-trace`, `github-repo`.
- - General fallback: `default`.
- If no prompt fits, proceed from `prompts/sources/default.md` and the user's
- brief.
- Style prompts under [`prompts/styles/`](./prompts/styles/) define reusable page
- systems such as `Timeline Story`, `Map Atlas`, `Network Map`, `Lesson Lab`,
- `Keepsake 3D Rhythm`, `Global Travel Map`,
- `Ops Console`, `Soft SaaS Console`, `Kami Longform Reader`, `Terminal CLI`,
- `Terminal Evidence Workbench`, `Mycelium Writing Environment` (`living-essay`),
- `Editorial Carousel`, and `Digital E-Guide Spread`. They complement source
- prompts; they do not replace source-specific analysis. The style prompt is
- binding for the final HTML's layout and interaction system.
README.md · README · 400 lines
View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)
- # html-anything
- > **The agent skill that turns anything into a beautiful, shareable HTML page.**
- > Rich answers, files, folders, URLs, and messy service exports become verified
- > single-file `.html` artifacts with source-aware parsing, automatic style
- > routing, **60 source prompts**, and **17 concrete style systems**. Short chats
- > stay short; page-worthy answers become pages.
- <p align="center">
- <a href="https://skills.sh/clockless-org/html-anything"><img alt="skills.sh" src="https://skills.sh/b/clockless-org/html-anything"></a>
- <a href="https://github.com/clockless-org/html-anything/stargazers"><img alt="GitHub stars" src="https://img.shields.io/github/stars/clockless-org/html-anything?style=flat-square"></a>
- <a href="https://github.com/clockless-org/html-anything/blob/main/LICENSE"><img alt="License" src="https://img.shields.io/github/license/clockless-org/html-anything?style=flat-square"></a>
- <a href="./prompts/sources"><img alt="60 source prompts" src="https://img.shields.io/badge/sources-60%20prompts-7b40e0?style=flat-square"></a>
- <a href="./prompts/styles"><img alt="17 style systems" src="https://img.shields.io/badge/styles-17%20systems-a03b00?style=flat-square"></a>
- <a href="https://clockless-org.github.io/html-anything/examples/"><img alt="11 demos" src="https://img.shields.io/badge/examples-11%20demos-0f766e?style=flat-square"></a>
- </p>
- <p align="center">
- <a href="https://clockless-org.github.io/html-anything/examples/"><strong>→ Open the live demo gallery</strong></a>
- </p>
- `html-anything` works inside Codex and Claude Code. Ask for a teaching site,
- drop in a PDF, upload a CSV, point it at a folder, or hand it an export like
- Amazon orders, Kindle highlights, Spotify history, WeChat / iMessage,
- Google Photos Takeout, logs, GPX, and more. The skill figures out the source,
- chooses the right use case and style, builds the page, checks it in a browser,
- and hands back a live HTML artifact instead of a long Markdown reply.
- ## Why this exists
- AI agents are outgrowing plain Markdown. Inspired by Claude Code team member
- Thariq Shihipar's viral
- [“The Unreasonable Effectiveness of HTML”](https://x.com/trq212/status/2052809885763747935),
- `html-anything` treats HTML as the richer response format: denser, more visual,
- more interactive, and easier to share when the answer is really a page.
- <table>
- <tr>
- <td width="100%" valign="top">
- <h3>Before: plain response</h3>
- <p>A long scroll of text for “Teach me solar system”. Useful, but hard to explore.</p>
- <img src="./docs/plain-response-before.jpg" alt="Plain AI response about the solar system" width="100%">
- </td>
- </tr>
- <tr>
- <td align="center">
- <strong>↓</strong><br>
- <sub>Same intent, richer response format</sub>
- </td>
- </tr>
- <tr>
- <td width="100%" valign="top">
- <h3>After: live HTML artifact</h3>
- <p>The same teaching goal becomes a visual, interactive learning page.</p>
- <a href="https://clockless-org.github.io/html-anything/examples/solar-system-studio/output.html"><img src="./docs/example-demos/solar-system-studio.gif" alt="html-anything solar system teaching studio (animated demo)" width="100%"></a>
- </td>
- </tr>
- </table>
- ## Preview
- <sub>Each card shows the kind of question a user would naturally ask. The skill detects intent, picks the right design system, and ships a polished HTML page — without anyone mentioning HTML. Open the [live demo gallery](https://clockless-org.github.io/html-anything/examples/) to see the rendered outputs.</sub>
- ### Teaching Studios
- <table>
- <tr>
- <td width="50%" valign="top">
- <a href="https://clockless-org.github.io/html-anything/examples/solar-system-studio/output.html"><img src="./docs/example-demos/solar-system-studio.gif" alt="Teach a concept demo" width="100%"></a><br>
- <strong><a href="https://clockless-org.github.io/html-anything/examples/solar-system-studio/output.html">Teach a concept →</a></strong><br>
- <sub>Source: teaching brief · Style: <code>teaching</code></sub><br>
- <sub><strong>Ask:</strong> <code>Teach me the solar system.</code></sub>
- </td>
- <td width="50%" valign="top">
- <a href="https://clockless-org.github.io/html-anything/examples/markdown/output.html"><img src="./docs/example-demos/markdown.gif" alt="Learn from long-form text demo" width="100%"></a><br>
- <strong><a href="https://clockless-org.github.io/html-anything/examples/markdown/output.html">Learn from long-form text →</a></strong><br>
- <sub>Source: Markdown file · Style: <code>architectural-spread</code></sub><br>
- <sub><strong>Ask:</strong> <code>Help me understand this essay.</code> (with a markdown file attached)</sub>
- </td>
- </tr>
- <tr>
- <td width="50%" valign="top">
- <a href="https://clockless-org.github.io/html-anything/examples/docx/output.html"><img src="./docs/example-demos/docx.gif" alt="Learn from a document demo" width="100%"></a><br>
- <strong><a href="https://clockless-org.github.io/html-anything/examples/docx/output.html">Learn from a document →</a></strong><br>
- <sub>Source: DOCX file · Style: <code>kami-reading</code></sub><br>
- <sub><strong>Ask:</strong> <code>Make this memo easier to read.</code> (with a docx file attached)</sub>
- </td>
- <td width="50%" valign="top"></td>
- </tr>
- </table>
- ### Files & Work Data
- <table>
- <tr>
- <td width="50%" valign="top">
- <a href="https://clockless-org.github.io/html-anything/examples/editorial-carousel/output.html"><img src="./docs/example-demos/editorial-carousel.gif" alt="Argument as sequence demo" width="100%"></a><br>
- <strong><a href="https://clockless-org.github.io/html-anything/examples/editorial-carousel/output.html">Argument as sequence →</a></strong><br>
- <sub>Source: strategy essay · Style: <code>editorial-carousel</code></sub><br>
- <sub><strong>Ask:</strong> <code>Help me share this strategy essay with my team.</code></sub>
- </td>
- <td width="50%" valign="top">
- <a href="https://clockless-org.github.io/html-anything/examples/pdf/output.html"><img src="./docs/example-demos/pdf.gif" alt="Guide from dense document demo" width="100%"></a><br>
- <strong><a href="https://clockless-org.github.io/html-anything/examples/pdf/output.html">Guide from dense document →</a></strong><br>
- <sub>Source: PDF report · Style: <code>digital-eguide</code></sub><br>
- <sub><strong>Ask:</strong> <code>Summarize this report and let me browse it section by section.</code></sub>
- </td>
- </tr>
- <tr>
- <td width="50%" valign="top">
- <a href="https://clockless-org.github.io/html-anything/examples/email/output.html"><img src="./docs/example-demos/email.gif" alt="Inbox or workstream audit demo" width="100%"></a><br>
- <strong><a href="https://clockless-org.github.io/html-anything/examples/email/output.html">Inbox or workstream audit →</a></strong><br>
- <sub>Source: Mbox archive · Style: <code>soft-saas</code></sub><br>
- <sub><strong>Ask:</strong> <code>What's happening in my inbox? Show me the open loops.</code></sub>
- </td>
- <td width="50%" valign="top">
- <a href="https://clockless-org.github.io/html-anything/examples/ci-log/output.html"><img src="./docs/example-demos/ci-log.gif" alt="Debugging evidence demo" width="100%"></a><br>
- <strong><a href="https://clockless-org.github.io/html-anything/examples/ci-log/output.html">Debugging evidence →</a></strong><br>
- <sub>Source: CI log · Style: <code>terminal-cli</code></sub><br>
- <sub><strong>Ask:</strong> <code>Why did this build fail? Walk me through it.</code></sub>
- </td>
- </tr>
- </table>
- ### Conversation Analysis
- <table>
- <tr>
- <td width="50%" valign="top">
- <a href="https://clockless-org.github.io/html-anything/examples/wechat-couple/output.html"><img src="./docs/example-demos/wechat-couple.gif" alt="Private chat recap demo" width="100%"></a><br>
- <strong><a href="https://clockless-org.github.io/html-anything/examples/wechat-couple/output.html">Private chat recap →</a></strong><br>
- <sub>Source: 1:1 chat export · Style: <code>love-romance-3d</code></sub><br>
- <sub><strong>Ask:</strong> <code>Recap our 1:1 chat history. Keep it private — mask names.</code></sub>
- </td>
- <td width="50%" valign="top">
- <a href="https://clockless-org.github.io/html-anything/examples/slack/output.html"><img src="./docs/example-demos/slack.gif" alt="Group contribution analysis demo" width="100%"></a><br>
- <strong><a href="https://clockless-org.github.io/html-anything/examples/slack/output.html">Group contribution analysis →</a></strong><br>
- <sub>Source: team chat export · Style: <code>kinetic-scoreboard</code></sub><br>
- <sub><strong>Ask:</strong> <code>Who's driving the conversation in this Slack channel?</code></sub>
- </td>
- </tr>
- </table>
- ### Personal Data & Places
- <table>
- <tr>
- <td width="50%" valign="top">
- <a href="https://clockless-org.github.io/html-anything/examples/kindle-highlights/output.html"><img src="./docs/example-demos/kindle-highlights.gif" alt="Reflective reading archive demo" width="100%"></a><br>
- <strong><a href="https://clockless-org.github.io/html-anything/examples/kindle-highlights/output.html">Reflective reading archive →</a></strong><br>
- <sub>Source: My Clippings.txt · Style: <code>living-essay</code></sub><br>
- <sub><strong>Ask:</strong> <code>What have I been reading and thinking about?</code></sub>
- </td>
- <td width="50%" valign="top">
- <a href="https://clockless-org.github.io/html-anything/examples/travel-history/output.html"><img src="./docs/example-demos/travel-history.gif" alt="Mobility recap demo" width="100%"></a><br>
- <strong><a href="https://clockless-org.github.io/html-anything/examples/travel-history/output.html">Mobility recap →</a></strong><br>
- <sub>Source: Uber/Lyft CSV · Style: <code>global-travel</code></sub><br>
- <sub><strong>Ask:</strong> <code>Where have I been this year? Show me on a map.</code></sub>
- </td>
- </tr>
- </table>
- ## Install
- Pick the path for your agent:
- [Claude Code](#claude-code) ·
- [Codex](#codex) ·
- [claude.ai](#claudeai-web) ·
- [Claude API](#claude-api) ·
- [Cursor / Cline / Windsurf / OpenCode / Goose / Letta / …](#one-command-for-most-cli-and-editor-agents) ·
- [ClawHub](#clawhub-publish)
- ### One command for most CLI and editor agents
- ```bash
- npx skills add clockless-org/html-anything
- ```
- Works with **Claude Code**, **Codex**, **Cursor**, **Cline**, **Windsurf**,
- **OpenCode**, **Goose**, **Letta**, **Roo Code**, **Kiro**, and any other
- agent following the [open agent-skills spec](https://agentskills.io).
- The CLI also pings [skills.sh](https://skills.sh/clockless-org/html-anything)
- so installs feed the public leaderboard.
- ### Claude Code
- ```bash
- git clone https://github.com/clockless-org/html-anything ~/.claude/skills/html-anything
- ```
- Restart Claude Code so it loads `SKILL.md`. To update later:
- `git -C ~/.claude/skills/html-anything pull`.
- ### Codex
- ```bash
- git clone https://github.com/clockless-org/html-anything "${CODEX_HOME:-$HOME/.codex}/skills/html-anything"
- ```
- Restart Codex.
- ### claude.ai (web)
- 1. Download [`html-anything-skill.zip`](https://github.com/clockless-org/html-anything/releases/latest/download/html-anything-skill.zip) (`SKILL.md` + `prompts/`, including style references).
- 2. In claude.ai: **Settings → Features → Skills → Upload a Skill** → drop the zip.
- Requires **Pro / Max / Team / Enterprise** with code execution enabled.
- Custom skills on claude.ai are per-user; each teammate uploads their own copy.
- ### Claude API
- ```bash
- curl https://api.anthropic.com/v1/skills \
- -H "x-api-key: $ANTHROPIC_API_KEY" \
- -H "anthropic-beta: skills-2025-10-02,code-execution-2025-08-25,files-api-2025-04-14" \
- -F file=@html-anything-skill.zip
- ```
- Then reference the returned `skill_id` in the `container` of any message.
- See the [Skills API guide](https://platform.claude.com/docs/en/build-with-claude/skills-guide).
- Uploaded skills are workspace-wide.
- ### ClawHub (publish)
- This repo carries `.clawhubignore` so only `SKILL.md` + `prompts/` ship,
- including canonical style references used by installed agents:
- ```bash
- npm i -g clawhub && clawhub login
- clawhub publish . --slug html-anything --version 0.1.1 --tags latest
- ```
- ### Where it does not work
- | Surface | Status |
- |---|---|
- | ChatGPT / chatgpt.com | ❌ Custom GPTs are a different format. Use **OpenAI Codex CLI** instead (above). |
- | Gemini web / Google Gems | ❌ Different format. Use **Gemini CLI** (`npx skills add`) instead. |
- | Anthropic / Claude Desktop | ✅ Reads `~/.claude/skills/` — same as Claude Code. |
- > ℹ Custom Skills do not sync across surfaces. The same skill uploaded to claude.ai is **not** automatically available on Claude Code or the API — each surface keeps its own copy.
- ## Use
- Just ask your real question. You do not need to mention HTML, pages, dashboards,
- reports, atlases, or any of the design vocabulary — the skill decides when an
- answer should be a page and ships one.
- ```text
- Teach me the solar system.
- What does my Amazon order history say about me?
- Help me understand this CSV.
- What did I read in 2025?
- Walk me through this GitHub repo.
- How did last quarter go?
- Make sense of this messy log file.
- Recap our 1:1 chat from this year.
- ```
- Each of these triggers a polished, single-file HTML page in the right design
- system. Short conversational asks stay short.
- If you name a data source but have no file yet (*"my Spotify history"*,
- *"my WhatsApp chat"*, *"my Google Photos Takeout"*), the skill walks you
- through the export first.
- ## Input And Output
- | Input | What you give | What you get |
- |---|---|---|
- | Rich answer | A topic, analysis request, comparison, recap, brief, or teaching goal | A readable, styled HTML artifact instead of a long written answer |
- | Idea | A short brief, e.g. "make a solar system teaching site" | A generated educational / interactive HTML page |
- | File | CSV, JSON, Markdown, PDF, DOCX, chat export, log, transcript, statement | A live page designed for that file |
- | Folder | Notes vault, Google Photos Takeout, Notion export, repo, exported archive | A browsable atlas / dashboard / audit page |
- | URL | Article, GitHub repo, public webpage | A shareable HTML reading or exploration page |
- | Export request | "My Amazon orders", "my Spotify history", "my relationship chat" | Export instructions first, then a live HTML page |
- The output is a browser page, not a chat reply. Most outputs are a single
- `output.html`. When the page needs generated images or other local
- assets, the skill returns `output.html + assets/`. Ask for "single-file"
- if you need everything in one HTML file.
- ## Automatic Usage Routing
- You do not need to choose a style. The default is `auto`.
- Routing has three layers:
- | Layer | Meaning |
- |---|---|
- | Use case | The user's job: teaching, files/work data, conversation analysis, or personal data/places |
- | Source | The input shape: prompt, CSV, PDF, DOCX, chat export, log, repo, folder, URL |
- | Style | The design system + layout system used to make the HTML readable |
- Styles are not CSS skins. The skill picks the system from the content, then
- builds the page inside that system. Every non-fallback style has a checked-in
- live example and screenshot preview.
- Style fidelity is part of the contract: when a style is based on a reference
- HTML or screenshot, the generated page should reproduce the reference's first
- viewport, component vocabulary, interaction model, motion grammar, and visual
- absence rules. Source modules are translated into the style instead of forcing
- every output into the same dashboard/report shape.
- Canonical style references can live under `prompts/styles/references/<style>/`,
- so the published skill can use the same structural target as the demo gallery
- without bundling every example. Style-specific assets stay beside their own
- reference pack and are copied only when that style needs them.
- | Usage pattern | Style |
- |---|---|
- | Unknown or mixed inputs | `default` (Insight Brief) |
- | Tutorials, lessons, explainers, "teach me" prompts | `teaching` (Lesson Lab) |
- | 1:1 chats and intimate message exports | `love-romance-3d` (Keepsake 3D Rhythm) |
- | Reflective essays, Kindle highlights, idea notes, concept-heavy reading archives | `living-essay` (Mycelium Writing Environment) |
- | Multi-participant activity streams, team chats, ranked contributors, owner/reps/players by workload | `kinetic-scoreboard` (Kinetic Championship) |
- | Personal histories — chronological (orders, history, listening, health) **and** topical (Notion / Obsidian vaults) | `timeline-story` (Timeline Story) |
- | Travel history, Uber/Lyft exports, and personal mobility recaps | `global-travel` (Global Travel Map) |
- | Places, trips, routes, geotagged photos | `map-atlas` (Map Atlas) |
- | Contacts, communities, social payments | `network-map` (Network Map) |
- | Support mailboxes, email campaigns, onboarding, customer-success queues | `soft-saas` (Soft SaaS Console) |
- | Finance, spreadsheets, logs, backlog, operational data | `dashboard` (Ops Console) |
- | Essays, articles, reading lists, bookmarks, PDFs, DOCX, legal/medical/lab records | `document` (Document Review) |
- | Long prose, DOCX memos, articles, essays, and manuscripts meant for sustained reading | `kami-reading` (Kami Longform Reader) |
- | Long-form visual explainers, object-focused articles, architectural split-screen editorial requests | `architectural-spread` (Architectural Editorial Spread) |
- | E-guides, PDF guides, creator guides, playbooks, lead magnets | `digital-eguide` (Digital E-Guide Spread) |
- | Brand strategy essays, founder letters, article takeaways, lightweight reports meant to be shared as a sequence | `editorial-carousel` (Editorial Carousel) |
- | Explicit terminal, CLI, shell, mainframe, hacker-console requests | `terminal-cli` (Terminal CLI, explicit override) |
- | Logs, PR patches, stack traces, CI failures, repos | `developer` (Terminal Evidence Workbench) |
- You can still steer it naturally: "make it more tutorial-like", "more
- app-like", "less academic", "make it a carousel", "more dashboard-like",
- or "more playful".
- Reusable style prompts live in [`prompts/styles/`](./prompts/styles/).
- The shared structural contract is
- [`prompts/styles/_system.md`](./prompts/styles/_system.md). The internal
- style catalog lives in [`prompts/styles/catalog.json`](./prompts/styles/catalog.json):
- it records the four use cases plus each style's triggers, best sources,
- example, preview, required primitives, and avoid rules so generation can stay
- style-faithful without asking users to pick options. There is a fallback
- `default` style plus 17 concrete style systems, each with a live example and
- preview asset.
- Example explicit style override:
- ```bash
- npx tsx src/cli.ts examples/pdf/input.pdf \
- --style digital-eguide \
- --out /tmp/battery-storage-guide.html \
- --title "Mid-Market Battery Storage Field Guide"
- ```
- ## Use Cases And Sources
- Sources can be endless, but the skill routes them into four stable use cases.
- Each use case can use one or more style systems.
- | Use case | Example sources | Likely styles |
- |---|---|---|
- | Teaching Studios | A short teaching brief, article, lesson outline, concept note, URL, Markdown, DOCX, or PDF/document simplification request | `teaching`, `architectural-spread`, `kami-reading` |
- | Files & Work Data | CSV / TSV, spreadsheet-style exports, JSON, JSONL, logs, CI output, PR patches, stack traces, repos, email/support archives, bank transactions, invoices, QuickBooks, calendars, issue trackers, Markdown, PDF, DOCX, bookmarks, URL lists, bibliographies, research records, slide-style carousel outputs | `dashboard`, `soft-saas`, `document`, `kami-reading`, `architectural-spread`, `digital-eguide`, `editorial-carousel`, `developer`, `terminal-cli` |
- | Conversation Analysis | WeChat, iMessage-style CSV, Slack, Discord, Telegram, email-style threads | `love-romance-3d`, `kinetic-scoreboard`, `network-map` |
- | Personal Data & Places | Amazon orders, Apple Health, browser history, YouTube, Spotify, Twitch, Kindle highlights, Venmo / PayPal, AI chat exports, notes vaults, Google Maps saved places, travel history, GPX, KML, itinerary CSV, location history | `timeline-story`, `global-travel`, `living-essay`, `network-map`, `map-atlas` |
- Use case is user-facing; style is internal. The user never has to name a
- style — natural asks like *"help me understand this CSV"* or *"teach me the
- solar system"* are enough; the skill picks the right system automatically.
- The detailed source-specific instructions live in [`prompts/sources/`](./prompts/sources/).
- ## Defaults
- - The skill chooses the style automatically.
- - The skill samples large sources, but renders the full data where practical.
- - The skill checks the page in a browser before handing it back.
- - Generated pages are local-first and static. They can be opened directly or hosted anywhere static HTML works.
- - Generated HTML can embed private source data client-side. Treat the output as sensitive as the original export.
- - Sensitive-record outputs are for organization and review only, not medical, legal, tax, accounting, immigration, insurance, or investment advice.
- ## Developer Note
- This repo also contains a standalone parser / CLI framework used by some
- examples, but the primary product surface is the agent skill. Users should
- not need to understand the internal implementation to use html-anything.
- ```bash
- git clone https://github.com/clockless-org/html-anything
- cd html-anything
- npm install
- export ANTHROPIC_API_KEY=sk-ant-... # or OPENAI_API_KEY=sk-...
- npx tsx src/cli.ts examples/csv/input.csv --out /tmp/customers.html
- ```
- ## License
- [MIT-0](./LICENSE)
Ratings and comments
Sign in to rate this skill.
Sign inFree to use. If it’s helped, pass it on.