CommunityVideo & Animationgithub.com

Azdmjiny/html-ppt-creator

Codex skill for creating polished, self-contained HTML slide presentations

Funktioniert mit~Claude CodeCodex CLI~Cursor
npx skills add Azdmjiny/html-ppt-creator

Ask in your favorite AI

Open a new chat with this agent skill pre-loaded.

Dokumentation

HTML PPT Creator

Create a clear, visually intentional HTML slide deck that opens directly in a browser. Default to one self-contained presentation.html with inline CSS and JavaScript and no build step.

Workflow

  1. Inspect the user's brief, source text, data, brand assets, and requested output location.
  2. Infer reasonable defaults for missing details. Ask only when audience, purpose, language, slide count, or visual direction would materially change the result and cannot be inferred.
  3. Design a concise slide sequence. Give each slide one job and a useful title; remove repetition before coding.
  4. Choose a visual system that fits the subject. Define CSS custom properties for color, type, spacing, borders, shadows, and motion, then use them consistently.
  5. Copy assets/starter.html as the base when starting from scratch. Replace all sample slides and theme tokens rather than treating the sample styling as a mandatory theme.
  6. Implement the complete deck. Use semantic <section class="slide"> elements, a fixed 16:9 stage that scales to the viewport, and keyboard/click/touch navigation.
  7. Read references/QUALITY.md before final QA. Open the deck in a browser when browser control is available, inspect every slide, fix overflow and interaction defects, then deliver the HTML file.

Do not introduce a planning checkpoint or require chapter-by-chapter approval unless the user requests an iterative workflow.

Output Contract

  • Produce presentation.html by default. Use a multi-file folder only when the user requests it or the deck genuinely needs reusable local assets or substantial application logic.
  • Keep the default deck offline-capable: inline CSS and JavaScript; use local or embedded images and fonts. Avoid CDN dependencies unless the user approves them.
  • Preserve supplied facts, wording constraints, citations, logos, and brand rules. Never invent statistics or sources.
  • Include a title slide, a coherent body, and a purposeful closing slide when appropriate; do not force these when the requested format says otherwise.
  • Support ArrowLeft, ArrowRight, Space, PageUp, PageDown, Home, and End; support click/tap navigation and touch swipes. Use F for fullscreen.
  • Show a subtle slide counter or progress indicator. Keep controls visually quiet and presentation-safe.
  • Add print CSS so each slide prints as one landscape page and decorative controls are hidden.
  • Respect prefers-reduced-motion. Use motion sparingly and never rely on animation to reveal essential information.
  • Do not add speaker notes, narration scripts, audio, autoplay, recording helpers, video timelines, or TTS infrastructure unless explicitly requested.

Content and Visual Rules

  • Make one main claim per slide. Prefer 3–6 concise supporting items over paragraphs.
  • Use the right visual form for the relationship: comparison, process, timeline, hierarchy, metric, quote, image, chart, or diagram.
  • Prefer real content, simple SVG, CSS diagrams, and meaningful charts over decorative filler.
  • Maintain strong hierarchy and generous whitespace. Design for viewing at a distance, not document reading.
  • Keep a coherent theme across the deck while allowing layout variety. Avoid making every slide the same grid of rounded cards.
  • Avoid generic AI styling: gratuitous purple gradients, excessive glow, random glass panels, emoji decoration, fake app chrome, and ornamental charts without data.
  • Ensure text contrast, visible focus states, useful alt text, and a sensible DOM reading order.
  • If content does not fit, shorten it or split the slide. Do not shrink body text until it technically fits.

Editing Existing Decks

Preserve the current architecture and design language unless the user asks for a redesign. Make the smallest coherent change, then recheck navigation, slide numbering, overflow, print layout, and responsive scaling.

Completion Check

Before reporting completion:

  1. Verify the requested messages, numbers, citations, and slide order.
  2. Test first/last-slide boundaries and every navigation method.
  3. Inspect all slides at 16:9 desktop size and at least one smaller viewport.
  4. Check for clipped text, collisions, tiny type, broken media, console errors, and accidental scrolling.
  5. Confirm the file opens without a build step and remains usable offline unless external dependencies were approved.
  6. Fix every failure before handing off the result.

Verwandte Skills

Ajaysinghadiya/LeadGen

Agentic lead generation system: finds local businesses without websites, generates AI site previews, records video tours, sends personalized WhatsApp outreach — orchestrated by Claude AI

community

agentspace-so/controlnet-pose

Pose-conditioned generation on RunComfy via the `runcomfy` CLI. Routes across Kling 2-6 Motion Control Pro / Standard (transfer the motion / blocking of a reference video onto a target character), community Wan 2-2 Animate (audio-driven character animation with pose conditioning), and Z-Image Turbo ControlNet LoRA (pose-conditioned image generation from an OpenPose / DWPose / canny / depth control image). Picks the right route based on video vs still and stylized vs photoreal. Triggers on "controlnet", "control net", "pose control", "openpose", "DWPose", "transfer pose", "motion control", "pose driven", "character pose", "depth control", "canny edge", "use this pose", or any explicit ask to condition generation on a pose / skeleton / motion / depth / canny reference.

community

hokeem/warm-drama-copy-skill

Codex skill for turning short-video references into copy scripts and adapted scripts

community

viper0355/claude-skills-pack

繁中在地化 Claude Code skills(youtube-watcher / humanizer / nano-banana-pro)+ API key 教學。推薦搭配 Tavily 官方 skills: github.com/tavily-ai/skills 與 Axton ai-pair 多 AI 協作: github.com/axtonliu/ai-pair

community

Nabil-MEDIOUNI/ytextract

🎬 Extract YouTube transcripts (single video or playlist) via a Playwright-driven CLI and AI-agent skill that drives YouTube's "Show transcript" panel.

community

greensock/gsap-timeline

Official GSAP skill for timelines — gsap.timeline(), position parameter, nesting, playback. Use when sequencing animations, choreographing keyframes, or when the user asks about animation sequencing, timelines, or animation order (in GSAP or when recommending a library that supports timelines).

community