agentleFS
maxedappsmaxedapps/agent-skills/skills/create-slides/SKILL.md

create-slides

Skill59 starsChanged 52 days agoMITView on GitHub

Creates and materially redesigns polished, dependency-free HTML slide decks from five styled templates, with stepped and automatic reveal animations driven by named motion presets, a cross-slide title morph, and PDF plus MP4 export at 1080p, 2K and 4K with reveals optionally timed to a narration/caption track. Use this skill when asked to create, build, redesign, restyle, or export a slide deck, presentation, pitch deck, talk, tutorial/video slides, course slides, slide b-roll for a video, or keynote-style slides as a web page. Do not use for ordinary websites or apps, editing proprietary slide files (PowerPoint, Keynote, Google Slides), research-only requests without a deck deliverable, or generating slide images only.

  • Installs packages

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. masseater/skills★ 2same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository8 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 3 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    No libraries, CDNs, frameworks, build steps, remote fonts, or network requests in the deck itself.

    line 29
  2. Distinctive:

    Never invent facts, statistics, quotes, images, or licenses.

    line 39
  3. Distinctive:

    Do not invent a new layout system mid-pass.

    line 43
  4. Very specific:

    Write semantic HTML: real headings, lists, figures, one <section class="slide"> per slide in DOM order.

    line 44
  5. Distinctive:

    User-owned copy is sacred: After the user edits text, change layout and CSS only — never rewrite their wording unless asked.

    line 51
  6. Very specific:

    create an empty deck.css

    line 75
  7. Very specific:

    Verify: npm run qa, and fix everything it reports.

    line 83
  8. Very specific:

    Do not add presenter consoles, autoplay, URL routing, or touch gestures unless asked.

    line 112

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Create Slides

Critical rules

  • No libraries, CDNs, frameworks, build steps, remote fonts, or network requests in the deck itself. It must open directly from a local index.html.
  • Never rebuild core/slides.js. Navigation, key handling and the data-enter/data-exit contract are fixed. Extend by layering on SlidesRuntime.controller, the way core/morph.js does. If the runtime itself must change, mirror it in scripts/slides-runtime.test.mjs.
  • Never edit core/slides.css per deck. It owns structure, archetype layout and the reveal presets. Override in theme.css, or in deck.css.
  • Fonts must be inlined as data URIs (tools/fonts.mjs). @font-face files are blocked over file://, so a deck opened by double-click loses them.
  • Never invent facts, statistics, quotes, images, or licenses. Verify current claims against primary sources, cite them, and record where every asset came from.
  • Derive tokens and content from the subject; derive the look from a template. Do not invent a new layout system mid-pass.
  • Write semantic HTML: real headings, lists, figures, one <section class="slide"> per slide in DOM order.
  • "Use the space" ≠ stretching boxes to fill the viewport. It means stable hierarchy, content-sized modules, comfortable gaps, and intentional margin.
  • Verify before reporting done. npm run qa for the deck, npm run audit after any recording. Never call a deck or an export finished on the strength of having looked at it.
  • User-owned copy is sacred. After the user edits text, change layout and CSS only — never rewrite their wording unless asked.

Workflow

  1. Inspect everything supplied — content, data, brand, delivery mode, and any script or captions the deck must back.
  2. Ask one grouped question covering only the real gaps. Always include the look unless the user named a template or gave brand direction that settles it. Defaults in brackets if you must proceed on assumptions:
    • Look: dark-marker (courses, tutorials) | light-editorial (talks, reports) | midnight-tech (developer content) | bold-keynote (pitches) | minimal-mono (research talks) — one line each, and say which you'd pick
    • Delivery: live talk | recorded video/tutorial | self-paced send
    • Density: sparse keynote | tutorial-comfortable
  3. Research unknown or current facts before writing any slide body.
  4. Outline the deck as titles only, each title the slide's takeaway rather than a topic label, and check that the sequence alone carries the message before writing any body copy.
  5. Scaffold the deck:
    • assets/core/ → core/, assets/skeleton.html → index.html
    • the chosen assets/templates/<name>/ → theme.css, fonts.json
    • assets/tools/ → tools/, assets/starter/* → deck root (gitignore → .gitignore)
    • create an empty deck.css
    • npm install && node tools/fonts.mjs
    • npm run dev renders every archetype the template offers — read it before writing slides.
  6. Adapt tokens only to the subject, then write the slides. Read references/web-slides.md for the DOM, composition and archetype contract, and references/motion.md before setting any reveal or preset.
  7. Verify: npm run qa, and fix everything it reports.
  8. Export if asked — references/export.md: npm run pdf, npm run record / record:2k / record:4k, then npm run audit. When the deck backs a recorded voiceover, time the reveals from the caption file in a sidecar, never in the markup.
  9. Report the files, how to open them, the controls, the template used, every assumption, and sources for any researched claim or asset.

Resources

PathRead when
references/web-slides.mdBefore editing markup — DOM/state contract, composition, archetypes, presets
references/templates.mdChoosing, adapting, or authoring a template; the token contract
references/motion.mdAdding any reveal, preset, morph, or animation
references/export.mdProducing a PDF or video, or timing reveals to narration
assets/optional/router.js — hash deep links; copy in only when asked (web-slides.md)
scripts/slides-runtime.test.mjsGuards the runtime contract; run after any change to it

Assets and provenance

  • Logos: take the vendor's own site icon first; aggregators go stale. Verify the mark is current, record every source in a CREDITS.md beside the files, and never draw an approximation of a real logo.
  • Light-on-dark app icons need a full-bleed tile (.tool__tile--dark), not a white one.

Constraints

  • Do not add presenter consoles, autoplay, URL routing, or touch gestures unless asked. When URL routing is asked for, copy assets/optional/router.js rather than writing one — the jump-without-replaying-a-morph handling is the hard part.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.