Slidewright logo

Slidewright

Community
tronghieu
slidewright

Build interactive presentation websites — slide decks projected to a room, controlled by one presenter. Use this skill WHENEVER the user wants to create, build, or design a presentation, slide deck, talk, or "bài thuyết trình / slide / trình chiếu", and also when they ask to add/edit slides, restyle a deck, scaffold a new presentation project, reuse their presentation setup in another repo, control slides from a phone (remote, QR code, "điều khiển slide bằng điện thoại"), or export slides to PDF — even if they don't say the word "slide" but clearly need projected talk visuals (e.g. "I'm giving a talk next week and need visuals", "dựng deck cho buổi chia sẻ", "làm mấy trang chiếu cho hội thảo"). Supports two tracks: a zero-build single HTML file, or a Vite + React project. Do NOT use it for printable documents, normal web apps, or dashboards meant for individual users to interact with.

Overview

Publishertronghieu
Repositoryagent-skills
Skill nameslidewright
Stars
98
Forks
32
Bundled files
16
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 16 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by tronghieu on GitHub. Read the source before you install it.

Installation

Install the Slidewright AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/tronghieu/agent-skills/tree/main/skills/slidewright
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/tronghieu/agent-skills.git /tmp/agent-skills
mkdir -p .claude/skills
cp -r /tmp/agent-skills/skills/slidewright .claude/skills/slidewright
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Slidewright in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Slidewright on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Slidewright is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Interactive presentation builder

Build slide decks that run as a website and are projected to a room. The deck is the visual half of a talk; the presenter does the talking. Two ways to build, one shared design language.

First, the mental model (read before writing any code)

There is one operator (the presenter) on one machine. Slides go on a big screen; the audience only looks — they never click, type, or interact. Everything below follows from this. Four failure modes to avoid:

  • Text too small. Slides are read from across a room, not on a laptop. Never use web-reading sizes (text-sm, 16px) or fixed Tailwind text classes (text-4xl). Always use fluid clamp() so text scales with the screen. Honour the typography floor (body ≈ 30–36px, never below 28px on a 1080p canvas). See references/design-system.md.
  • Content hugging an edge. Padding is a two-layer contract. Slide content must stay inside the viewport safe area, and content inside a card, panel, callout, bordered box, or other visible container must stay inset from that container's own edge. A parent slide's padding does not protect a nested container; gap and child margins are not a substitute for the container's own padding. Both scaffolds provide safe-area padding and a .slide-surface inset helper. Even fullBleed slides keep text content inset.
  • Building an app. No input fields, no "Submit", no login, no data collection — there is no backend and nowhere for data to go. Interaction is only the presenter clicking to reveal/advance content. If a component asks "where does this data go?", it's wrong.
  • Wrong voice for the audience. A board report is not a lesson and a pitch is not a status update. The deck type sets the structure and the voice. Decide it first — see references/deck-types.md.

The presenter may hold a phone remote (scan a QR code from the terminal, then tap Next/Prev). It is the same single operator on a second device, not audience interaction.

references/design-system.md is the heart of this skill — the typography floor, two-layer padding contract, content overflow, layout recipes, motion, and palette. Read it whenever you design or restyle slide content.

Workflow

  1. Understand and clarify the talk.
    • Content: If the presentation content is vague or unclear, ask questions to clarify the topic, audience, speaker (name/role), and key message before building. For Vietnamese content, the default voice is plain, honest, direct — no marketing hype.
    • Deck type: Decide what the deck is for before writing: report upward, decision request, pitch, training, talk, or internal update. Infer it from the request and tell the user your choice in one line; ask only when two types fit equally. Then read references/deck-types.md and follow that type's arc, voice, and prohibitions.
    • Style: Ask the user for their preferred visual style and suggest these 5 modern options:
      1. Minimalist Bauhaus: High contrast, geometric typography, generous whitespace.
      2. Dark Mode Cinematic: Deep slate/black backgrounds, dramatic contrast, glowing accents.
      3. Editorial / Magazine: Sophisticated font pairings (serif + sans), structured grid layouts.
      4. Glassmorphism: Soft background gradients with blurred, translucent content panels.
      5. Neo-Brutalism: Bold typography, raw borders, hard shadows, high energy.
  2. Pick a track (below). If unsure, ask once; otherwise default to plain HTML for small/quick decks and React for substantial, maintained talks.
  3. Scaffold with the matching script — don't hand-assemble the boilerplate.
  4. Build slides in the arc and voice of the deck type, following references/design-system.md: one idea per slide, few words, real visuals, projection-legible type, presenter-only interaction, and the two-layer padding contract for both viewport and content containers. Respect the density limits — never exceed 5 bullets, ~60 words, or 5×4 table cells per slide. If content doesn't fit at the typography floor, split into more slides — never shrink past the floor and never produce a slide that scrolls vertically.
  5. Keep the required chrome: a visible bottom navigation slider (dot strip) and slide number. Both scaffolds include it — don't remove it.
  6. Write speaker notes in the <deck-name>-notes.md file the scaffold creates (never on the slides themselves).
  7. Run the verify script to catch text-too-small and content-too-dense violations:
    bash
    bash /mnt/skills/user/slidewright/scripts/verify-slides.sh <path-to-deck>
    Fix every reported issue before proceeding.
  8. Run the spacing audit before handoff. Check every slide at 1920×1080 and at one smaller viewport. No text or primary content may cross the viewport safe area. For every element that draws a visible boundary (background, border, outline, shadow, rounded surface, or text-overlay panel), verify that its content has padding on all four sides. Media and purely decorative edge-to-edge layers are the only exceptions; text over them belongs in a nested padded surface.
  9. Tell the user how to present with the phone remote. Both scaffolds include it. The HTML track runs node <deck>/slidewright-remote.mjs. The React track runs npm run dev. The terminal prints a QR code for the phone. See references/remote-control.md. Keep the deck's window.slidewright API and slidewright:change event when you edit navigation code.
  10. Export to PDF if asked — see references/export-pdf.md.

Put each deck in its own folder. Keep any live-demo app as a separate project, not inside the deck.

Choosing a track

PickWhenReference
Plain HTMLZero setup wanted; runs by opening a file; quick deck; shareable anywhere; no toolchain. Gets unwieldy past ~15–20 rich slides.references/html-track.md
Vite + ReactMany slides, reusable components, state-driven interactions, TypeScript, Framer Motion, maintained/re-run talks.references/react-track.md

Both produce the same projected experience and obey the same design system; they differ only in implementation.

Scaffolding

Plain HTML — single self-contained index.html, Tailwind via CDN, no build:

bash
bash scripts/new-html-deck.sh <deck-name> [target-dir] [--title "Deck title"]

Vite + React — scaffolds with the official Vite tool and installs the latest React, Tailwind, Framer Motion, and Lucide (versions are intentionally not pinned, so each deck starts on current tooling), then layers the App → Deck → Slide architecture:

bash
bash scripts/new-react-deck.sh <deck-name> [target-dir] [--no-install]
cd <deck-name> && npm run dev

Both scaffolds include the phone remote (slidewright-remote.mjs, zero dependencies). For a deck scaffolded before the remote existed:

bash
bash scripts/add-remote.sh <path-to-deck>

After scaffolding, read the matching track reference for how slides are structured and how to add/reorder them, then build the content.

Files in this skill

  • references/design-system.md — projection rules, typography floor, layout recipes, motion, palette. The core; consult for any content/design work.
  • references/deck-types.md — the six deck types and each one's arc, voice, and prohibitions. Read it before writing slide content.
  • references/html-track.md — plain-HTML deck structure, template internals, adding slides.
  • references/react-track.md — Vite+React architecture, slide ordering, Tailwind wiring.
  • references/export-pdf.md — PDF export options and speaker-notes convention.
  • references/remote-control.md — phone remote: running it, the deck API contract, security, troubleshooting (wrong IP, firewall, venue Wi-Fi).
  • scripts/new-html-deck.sh — scaffold a plain-HTML deck.
  • scripts/new-react-deck.sh — scaffold a Vite + React deck.
  • scripts/add-remote.sh — add the phone remote to an existing deck.
  • assets/remote/slidewright-remote.mjs — the phone remote: WebSocket relay, remote page, terminal QR code, static server (HTML track) and Vite plugin (React track). The scaffolds copy it into each deck; edit this source, not the copies.
  • scripts/export-deck-pdf.py — export a deck to a content-complete PDF (waits for render and reveals hidden content; image-based).
  • scripts/verify-slides.sh — static analysis of deck files: flags text below the typography floor, fixed-px font sizes, excessive bullet counts, and high word density per slide. Run after building slides and before handoff.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Slidewright AI skill do?

Build interactive presentation websites — slide decks projected to a room, controlled by one presenter. Use this skill WHENEVER the user wants to create, build, or design a presentation, slide deck, talk, or "bài thuyết trình / slide / trình chiếu", and also when they ask to add/edit slides, restyle a deck, scaffold a new presentation project, reuse their presentation setup in another repo, control slides from a phone (remote, QR code, "điều khiển slide bằng điện thoại"), or export slides to PDF — even if they don't say the word "slide" but clearly need projected talk visuals (e.g. "I'm giv...

Why use Slidewright on TypingMind?

Because you install it once and use it with any model. Slidewright is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Slidewright in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/tronghieu/agent-skills/tree/main/skills/slidewright. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Slidewright?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Slidewright?

As many as you like. As long as a model supports skills, you can use Slidewright with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Slidewright AI skill free?

Yes. It is published on GitHub by tronghieu under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇