What does this skill do, and when should you use it?
Guizang PPT Skill is an Agent Skill for Claude Code, Codex, and similar environments that generates horizontal, swipe-based HTML decks. It includes two visual systems: editorial magazine/e-ink and Swiss Internationalist. The output is a self-contained HTML presentation that opens directly in a browser and supports keyboard, wheel, touch, and index navigation. Swiss-style decks also use named layouts and a validation script.
Reads supplied articles, Markdown, data, or screenshot requirements; selects a visual style, theme, and layout based on the presentation brief; copies assets/template.html or assets/template-swiss.html; produces a single index.html containing the presentation structure, styles, navigation, WebGL or canvas effects, and motion behavior; references assets from an images/ directory using standard aspect ratios; runs scripts/validate-swiss-deck.mjs for Swiss-style decks; and can optionally use image generation in Codex to create deck artwork and covers.
- A writer or researcher turning a long article into a 6–10 page talk.
- A product team preparing an AI launch, demo day, or business presentation.
- A speaker expressing industry observations, cultural stories, or personal viewpoints with an editorial layout.
- A design, engineering, or product team presenting analysis, systems, metrics, or an annual review in a grid-based Swiss style.
- A content team extending one visual system into an official-account header, square share card, or Xiaohongshu cover.
How do you install this skill?
- The templates, validator, and browser preview were not executed, so actual HTML runtime behavior, layout correctness, and low-power behavior remain unverified.
- The core path may depend on CDNs, GitHub, npm, browser capabilities, and optional image-generation services; network reachability and offline fallbacks should be confirmed.
- Although the documentation mentions masking sensitive information, it does not provide a systematic privacy review, permission-control, or cleanup mechanism for screenshots and source materials.
- Installation instructions include git clone, npx, and writes to a user skill directory; the user should confirm the target path and network-dependent actions first.
- AGPL-3.0 obligations should be reviewed before redistribution or modification; sponsor and provenance statements do not constitute verification by the FollowSkills enterprise registry.
- Shell / CLI
- Network access
- Local filesystem
GitNode.js
Recommended command: npx skills add https://github.com/op7418/guizang-ppt-skill --skill guizang-ppt-skill. Alternatively, clone it to ~/.claude/skills/guizang-ppt-skill with git clone https://github.com/op7418/guizang-ppt-skill.git ~/.claude/skills/guizang-ppt-skill. The installation should contain SKILL.md, assets/, and references/.
git clone --depth 1 https://github.com/op7418/guizang-ppt-skill.git ~/.claude/skills/guizang-ppt-skillGenerated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- Create an approximately 8-page Swiss-style presentation from this article, with three illustrations whose aspect ratios match the template slots.
After installation, give the Agent a prompt such as: “Create an approximately 8-page Swiss-style presentation from this article, with three illustrations whose aspect ratios match the template slots.” The documented workflow covers style selection, clarification of audience and assets, template copying, layout-based content filling, optional image generation, validation, and opening index.html in a browser. Before starting, the Skill checks the GitHub upstream for updates but does not update automatically.
What are this skill's strengths and limitations?
- Delivers a single HTML file that can be opened without a build step or local server.
- Provides both an editorial narrative system and an information-driven Swiss system.
- Includes 10 Style A layouts and 22 named Style B layouts.
- Supports keyboard, wheel, touch, ESC index navigation, and a B-key low-power static mode.
- Provides Swiss-style checks for layouts, image slots, and selected visual constraints.
- The primary deliverable is HTML rather than native PPTX; PPTX conversion is outside the documented workflow.
- It is not intended for dense tables, high-density training materials, or collaborative editing.
- The workflow depends on local shell access, filesystem access, a browser, and Node.js-based validation scripts.
- The source does not provide a comprehensive automated test suite or independent results across Agent clients.
- Image generation is optional, and the repository does not specify one mandatory image service or its cost.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Guizang Web Deck Skill this page | 53 · Use with care | ★ 28k | 2mo ago | AGPL-3.0 |
| Slides Grab Design Skill | 51 · Use with care | ★ 1.2k | 1mo ago | MIT |
| Slides Grab HTML Skill | 60 · Recommended | ★ 1.2k | 1mo ago | MIT |
| baoyu-design: Run Claude Design Locally as an Agent Skill | 53 · Use with care | ★ 4.4k | 1mo ago | MIT |
| HTML Anything: Turn Anything Into a Page | 53 · Use with care | ★ 156 | 5mo ago | MIT-0 |
Compared with conventional presentation software, Guizang PPT Skill delivers HTML/CSS that is easier for an Agent to generate and revise and can be presented directly in a browser, but it offers less native collaboration and PPTX support. Compared with Markdown, it supports finer typography, spatial layouts, animation, interaction, and cover formats.
How did FollowSkills review this skill?
The documentation requires clarification, mentions masking sensitive screenshot data, provides low-power behavior and local fallbacks, and discloses use of CDNs, images, and optional image generation; however, permission boundaries, third-party dependency security, network requests, privacy handling, user confirmation, rollback, and isolation of generated content are not sufficiently specified, so points are deducted.
The workflow includes template selection, class preflight checks, static validation, visual review, and some abnormal-case guidance; however, this review is static and provides no verifiable execution results, comprehensive test suite, or CI evidence. Browser, font, CDN, and image-generation failure diagnosis is also limited, so the static cap applies.
Audience, use cases, non-fit cases, trigger phrases, clarification inputs, and Chinese font fallbacks are relatively clear; however, the skill assumes filesystem, shell, browser, and agent capabilities, while reachability and fallback boundaries for CDN, GitHub, npm, and optional image services are not adequately documented, so points are deducted.
SKILL.md and README.en.md provide readable structure, installation, workflow, layouts, resources, FAQ, roadmap, contribution guidance, and an AGPL-3.0 license; however, there is no clear versioning scheme, changelog, maintenance responsibility, or stable update path, and some claimed resources and validation capabilities cannot be confirmed from the supplied evidence, so points are deducted.
The claimed output is a single-file HTML presentation, supported by two templates, layout skeletons, image-ratio guidance, and review steps, making the core value plausible; however, no representative artifact, test output, or comparative evidence is statically verifiable, and delivery still depends on manual visual QA, so the static cap requires a conservative score.
The supplied files provide auditable source material, installation commands, layout rules, and a validator path, while README sections distinguish some capabilities and limitations; however, there is no CI result, committed test evidence, independent reproduction record, or corroboration, leaving key effectiveness claims primarily author-reported, so points are deducted.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →