Index of every page on this site: llms.txt
All skills
Run it yourself

Scaffold, author, style, preview, and deploy dark-theme Slidev presentation decks with modular act files, custom Vue layouts, and embedded asciinema terminal recordings. Use when creating a new Slidev deck, authoring or refactoring slides and presenter notes, embedding .cast recordings or visual callouts, running or managing the local Slidev dev server, or configuring GitHub Pages deployment.

Use this Skill: https://skilld.dev/gh/domesticmouse/skills/slidev-deck

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈102 tokens for metadata: the name and description. ≈3.1k when used: this file. ≈4.8k more on demand in 7 files.

Description uses 5.1% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈409 characters, or 5.1%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

Slidev Deck

Procedural runbook, reusable Vue components/layouts, and server tooling for building high-signal technical slide decks with Slidev (slides/), embedded asciinema-player terminal recordings, dark-mode typography, and static GitHub Pages deployment.

Works alongside /asciinema (recording, sanitizing, and trimming .cast files), /ios-simulator-screenshot (staging and verifying mobile simulator screen captures under artifacts/screenshots/<project_name>/), and /crop-image (trimming transparent margins and solid borders on slide images).

Workflows

Choose the matching workflow based on your objective:

  1. Scaffold a New Slidev Deck
  2. Structure Modular Acts
  3. Author Slides, Layouts & Presenter Notes
  4. Embed Terminal Recordings & Visual Assets
  5. Run & Manage the Local Dev Server
  6. Verify Build & Deploy to GitHub Pages

1. Scaffold a New Slidev Deck

Initialize a new slides/ workspace from the bundled templates in resources/:

  1. Create the directory hierarchy and copy the starter configuration, theme stylesheet, component, and custom layouts:
    mkdir -p slides/{acts,components,layouts,public/casts,public/images}
    cp .agents/skills/slidev-deck/resources/package.json slides/package.json
    cp .agents/skills/slidev-deck/resources/vite.config.ts slides/vite.config.ts
    cp .agents/skills/slidev-deck/resources/style.css slides/style.css
    cp .agents/skills/slidev-deck/resources/slides.md slides/slides.md
    cp .agents/skills/slidev-deck/resources/acts/01-primer.md slides/acts/01-primer.md
    cp .agents/skills/slidev-deck/resources/components/AsciinemaPlayer.vue slides/components/AsciinemaPlayer.vue
    cp .agents/skills/slidev-deck/resources/layouts/*.vue slides/layouts/
  2. Update the --base /<repo-name>/ flag in slides/package.json ("build": "slidev build --base /<repo-name>/") to match the repository name for GitHub Pages subpath hosting (or omit --base if deploying to a custom root domain).
  3. Add slides/node_modules/, slides/dist/, slides/.slidev.pid, and slides/.slidev.log to .gitignore.
  4. Install dependencies:
    npm --prefix slides install
  5. Completion Criterion: slides/node_modules/ is populated and npm --prefix slides run build exits with code 0.

2. Structure Modular Acts

Split multi-part presentations into Act-level markdown files under slides/acts/ while keeping slides/slides.md as the master manifest:

  1. Configure slides/slides.md with global presentation headmatter (colorSchema: dark, background: '#121317', highlighter: shiki, mdc: true) and set hide: true on the headmatter block so Slidev does not insert a blank initial slide or leak headmatter fields into the first act:
    ---
    theme: default
    background: '#121317'
    colorSchema: dark
    highlighter: shiki
    lineNumbers: false
    info: |
      ## Presentation Title
      Subtitle and Presenter Info
    drawings:
      persist: false
    transition: slide-left
    title: Presentation Title
    mdc: true
    hide: true
    ---
    
    ---
    src: ./acts/01-primer.md
    ---
    
    ---
    src: ./acts/02-core-story.md
    ---
  2. Name act files sequentially (01-primer.md, 02-the-hook.md, etc.) inside slides/acts/.
  3. Tag every slide with a unique comment identifier immediately after its frontmatter block (e.g., <!-- S01 -->, <!-- S12 -->) to anchor edits and reviews.
  4. Completion Criterion: slides/slides.md contains hide: true in its headmatter block and every src: ./acts/... entry resolves to a valid markdown file in slides/acts/.

3. Author Slides, Layouts & Presenter Notes

Follow the slide design conventions below. For copy-pasteable HTML/Markdown templates of every slide archetype, read references/SLIDE-PATTERNS.md.

Core Formatting Conventions:
  1. Clean Slide Headings: Render slide titles using a clean flex header container without badges, pills, or Act chips:
    <div class="flex items-center gap-3 mb-6">
      <h2 class="text-2xl font-bold m-0">Slide Title</h2>
    </div>
  2. Numbered Concept Lists: Present multi-point explanations and comparisons as structured numbered lists (<ol>) with bold white titles and nested <ul> takeaways rather than cramped multi-card grids or dense multi-column tables:
    <ol class="space-y-4 text-sm text-gray-200 list-decimal pl-6 mt-5 leading-relaxed">
      <li>
        <strong class="text-white">Topic / Feature Title</strong>
        <ul class="list-disc pl-5 mt-1 space-y-1 text-gray-300">
          <li>Primary high-signal takeaway or mechanism</li>
          <li>Secondary context or contrast</li>
        </ul>
      </li>
    </ol>
    When comparing 3–4 broad paradigms (e.g., MCP vs. Rules vs. Skills), give each concept its own focused numbered-list slide followed by a synthesis slide rather than squeezing all columns into one tiny table.
  3. Flush Code Fences Inside HTML: Place markdown fenced code blocks (```) flush at column 0 with blank lines above and below whenever they sit inside HTML <div> wrappers. Indenting a fenced code block by 4+ spaces inside HTML causes Markdown to render raw tags instead of highlighted code.
  4. Conversational Presenter Notes: Conclude every slide with an HTML comment (<!-- ... -->) containing natural spoken narrative that explains the transition and talking points behind each bullet. Write pure spoken prose without bracketed stage directions ([Click], [Next slide]).
Choosing a Slide Layout:
  • layout: center (class: text-center): Title slide, Act divider slides, and big-quote callouts.
  • layout: default: Concept lists, 2-column anti-pattern vs. solution cards (bg-red-950/15 vs. bg-emerald-950/15), and code/architecture tables.
  • layout: cast-full: Full-viewport interactive asciinema-player slide (title, cast: /casts/file.cast, autoPlay: false, speed: 2.0, idleTimeLimit: 2.0, fit: both, and optional showChapterBar: true during development).
  • layout: skill-breakdown: Framed zoom-in card (way: 1|2|3|mashup, title).
  • layout: hardware-callout: 12-column split pairing a framed image and optional caption overlay on the left (title, image: /images/file.png, imageCaption) with explanatory HTML content in the default slot on the right.

Completion Criterion: Every authored slide includes a clean <h2> header (or uses a dedicated layout header), unindented code fences with blank line separators, and a conversational <!-- ... --> speaker note block.


4. Embed Terminal Recordings & Visual Assets

Consult references/ASCIINEMA-INTEGRATION.md for component props, Vite SSR rules, and base-path normalization details.

  1. Prepare & Embed .cast Recordings:
    • Before executing a multi-cast Recording Run, use /asciinema (.agents/skills/asciinema/references/RUNBOOK-FORMAT.md) to generate an operator RUNSHEET.md from TODO.md or the slide blueprint (separating recordings/raw/ from recordings/final/ at 120x36 with paired Before / After "m" stop points).
    • Treat recordings/final/<name>.cast as the single source of truth. Use /asciinema to sanitize secrets (sanitize-cast.py), strip trailing shell exit prompts (trim-cast.py --in-place), and place/verify "m" chapter stop points (mark-cast.py --list / --set) on the canonical recording in recordings/final/, then run npm --prefix slides run sync-casts (also executed automatically by predev and prebuild) to sync recordings/final/*.cast into slides/public/casts/.
    • For shell commands and TUI launchers in zsh (e.g. flutter create ..., git init && ..., agy), place command-entry stop points on the typed/pasted command text rather than post-launch banners; mark-cast.py automatically advances past bracketed-paste reverse video ([REV]) to the un-reversed redraw ([UNREV]) and splits trailing \x1b[1B\r\x1b[K cursor-down sequences into a post-marker "o" event (dt=0.02s) so the cursor stays on the same line at the end of the command.
    • Reference recordings with root-relative paths in slide frontmatter:
      ---
      layout: cast-full
      title: "Recording Title"
      cast: /casts/demo.cast
      autoPlay: false
      speed: 2.0
      idleTimeLimit: 2.0
      fit: both
      ---
    • AsciinemaPlayer.vue compiles each breakpoint-bounded segment to run at >= speed (default 2.0x) and complete in <= maxSegmentDuration (default 30s). By default in final presentation slides (showChapterBar: false), it renders a clean minimalist Playback speed: Nx badge beneath the terminal while supporting [ / ] / p stage hotkeys; set showChapterBar: true in slide frontmatter during development to display interactive chapter stepper pills, transport controls, and segment metadata.
  2. Prepare & Embed Images:
    • For mobile app runs, use /ios-simulator-screenshot (xcrun simctl io booted screenshot /abs/path/to/repo/artifacts/screenshots/<project_name>/<file>.png with full absolute paths) to stage clean device-framebuffer screenshots per run in artifacts/screenshots/<project_name>/.
    • Use /crop-image (crop.py --inspect / --autocrop) to verify dimensions or remove excess transparent margins/solid borders before copying presentation-ready images into slides/public/images/.
    • Reference images via image: /images/<filename> in hardware-callout slides.
  3. Preserve BASE_URL Resolution in Custom Vue Files:
    • In any custom Vue layout or component that loads local assets from slides/public/, normalize paths with import.meta.env.BASE_URL so assets resolve on both local dev (/) and GitHub Pages subpaths (/<repo-name>/).
  4. Completion Criterion: All /casts/* and /images/* paths referenced in slides/acts/*.md exist in slides/public/{casts,images}/.

5. Run & Manage the Local Dev Server

Manage the background Slidev development server using the bundled scripts/slides-server.sh utility:

# Start server on default port 3030 (auto-installs node_modules if missing)
.agents/skills/slidev-deck/scripts/slides-server.sh start

# Check server PID, HTTP status, and navigation URLs
.agents/skills/slidev-deck/scripts/slides-server.sh status

# Restart or stop the server
.agents/skills/slidev-deck/scripts/slides-server.sh restart
.agents/skills/slidev-deck/scripts/slides-server.sh stop

# Use a custom port if 3030 is in use
.agents/skills/slidev-deck/scripts/slides-server.sh start -p 3040
Live Slidev Endpoints:
  • Presentation Deck: http://localhost:3030/
  • Presenter Console (speaker notes, timer, next slide): http://localhost:3030/presenter
  • Slide Overview Grid: http://localhost:3030/overview

Completion Criterion: Running .agents/skills/slidev-deck/scripts/slides-server.sh status reports RUNNING with HTTP: 200 (or 304).


6. Verify Build & Deploy to GitHub Pages

  1. Run the static production build to verify that all Vue components, layouts, Shiki code blocks, and Markdown slides compile cleanly:
    npm --prefix slides run build
  2. To configure automated GitHub Pages deployment, copy the workflow template into .github/workflows/deploy-slides.yml:
    mkdir -p .github/workflows
    cp .agents/skills/slidev-deck/resources/deploy-slides.yml .github/workflows/deploy-slides.yml
    Ensure on.push.paths in .github/workflows/deploy-slides.yml includes every upstream directory synced by prebuild (e.g., 'slides/**', 'recordings/final/**', and any external tutorial or asset folders) so pushing changes to those files triggers the GitHub Pages workflow.
  3. Ensure repository GitHub Pages settings are configured with Build and deployment > Source: GitHub Actions.
  4. Completion Criterion: npm --prefix slides run build exits with returncode 0 and generates slides/dist/index.html.

Source: SKILL.md on GitHub

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No alerts16d3 checks · Risk SAFE
  • Gen Agent Trust Hub16d

    The Slidev Deck skill provides a robust framework for creating technical presentations with terminal recordings. It automates environment setup, dependency management, and local server control using standard Node.js tools. The skill was found to be safe, following best practices for local development workflows.

  • Socket16d

    No alerts

  • Snyk16d

    Risk: LOW · No issues

Signed by skilld at 606a5d5. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub yesterday.

Activeupdated 2 days ago
Other metadata
compatibility
Requires Node.js (v18+) and npm; integrates with Slidev (@slidev/cli) and asciinema-player v3

README badge

README badge for domesticmouse/skills/slidev-deck