HTML PPT Studio — AgentSkill with 24 themes, 31 layouts, 20+ animations for building professional HTML presentations
AI Analysis
HTML PPT Studio is a specialized AgentSkill for generating professional HTML presentations with 36 themes, 31 layouts, and 47 animations, featuring a unique presenter mode with speaker scripts and live previews. It serves AI agents and presentation developers who need to programmatically create pixel-perfect slides without a build step. This tool is not a general-purpose presentation editor but rather a framework for agents to author decks or for developers building presentation automation sy...
Inferred from signals mentioned in the README (tests, CI, type safety) — not a review of the actual code.
AI's overall editorial judgment — not an average of the bars above, can weigh other factors too.
HTML-only presentation toolkit for AI agents: 36 themes, 31 layouts, no build step
html-ppt-skill is a static HTML/CSS/JS presentation library packaged as an AgentSkill — a plugin format designed for AI agent runtimes. It lets agents generate professional-looking slide decks on demand without any build toolchain. Target users appear to be developers running local or hosted AI agents who want structured, themed presentation output. The presenter mode with BroadcastChannel sync is a notable feature. Its value lies in reducing the templating burden on agents producing slide content.
Created in April 2026, during a wave of similar AgentSkill repositories targeting the same AI-agent-driven presentation niche. The ecosystem of similar repos (guizang-ppt-skill, codex-ppt-skill) suggests this is a recognized pattern, not an isolated experiment.
The repo gained 6,372 stars since mid-April 2026 — roughly 10 weeks — suggesting strong initial interest from the Chinese-speaking developer community (evident from bilingual README and xiaohongshu templates). The 106 stars gained in the last 7 days indicates growth is slowing relative to likely early viral spikes. The 590 forks suggest active downstream customization rather than passive interest. Growth appears driven by the AI-agent tooling trend rather than purely organic word-of-mouth.
Adoption not verified via documented production deployments. Star and fork counts suggest developer interest, but there is no evidence of enterprise use, integration into major agent platforms, or third-party packages depending on this project. The AgentSkill install command (`npx skills add`) suggests it targets a specific agent runtime whose own adoption level is also not independently verifiable here.
Appears to be fully static HTML/CSS/JS with no build step. Likely uses CSS custom properties (tokens) per theme file, with a shared layout system and per-deck scoped class namespaces (`.tpl-<name>`) to prevent collisions. Presenter mode appears to use BroadcastChannel and postMessage for cross-window communication. Canvas FX animations likely rely on vanilla JS canvas elements. Architecture appears modular but not verified beyond README description.
README mentions 56 verification screenshots in `scripts/verify-output/` suggesting visual regression or smoke-test output exists, but no automated test suite or CI configuration is documented. The nature and reliability of these verification artifacts is unclear.
Last push was 2026-04-26 — approximately 8 weeks before the evaluation date of 2026-06-23. No pushes in nearly 2 months is a moderate concern for a project this young. Either the author considers it feature-complete at this stage, or active development has paused. No issue activity data is available to assess responsiveness.
ADOPT IF: you are building or using an AI agent that needs structured HTML slide output and your runtime supports the AgentSkill format — this offers a ready-made theme and layout library that reduces prompt engineering for visual consistency. AVOID IF: you need a maintained, production-hardened solution with active issue resolution, or if your agent runtime does not support the AgentSkill install mechanism. MONITOR IF: you are evaluating the AgentSkill ecosystem broadly — this repo's 2-month commit gap makes it unclear whether development will resume or stall.
Independent dimensions
Mainstream potential
3/10
Technical importance
4/10
Adoption evidence
2/10
- Commit activity has been dormant for approximately 8 weeks as of evaluation date; unclear whether this reflects completion or abandonment.
- The AgentSkill format itself is ecosystem-specific — if the target agent runtime loses traction, this tooling loses its primary distribution advantage.
- README contains numerical inconsistencies (headline claims 36 themes and 47 animations, description block lists matching numbers, but earlier project description says 24 themes and 20+ animations) — suggests documentation may not always match actual file contents.
- No documented test suite beyond static screenshots; correctness of theme/layout rendering across agent-generated content is not independently verifiable.
- Strong apparent focus on Chinese-language use cases (xiaohongshu templates, bilingual docs) may limit organic adoption growth in English-dominant developer communities.
Likely to remain a useful reference repo for the AI-agent presentation niche, but may plateau in adoption unless the author resumes active development or the underlying AgentSkill runtime gains wider platform support.
Newsletter
Get analyses like this every Monday
Free weekly digest of the most interesting open-source discoveries.
Languages
Information
- Language
- HTML
- License
- MIT
- Last updated
- 3mo ago
- Created
- 3mo ago
- Analyzed with
- anthropic/claude-haiku-4-5
Stars over time
Contributors over time
Top 100 contributors only — repos with more will plateau at 100.
Open issues
Similar repos
GordenSun/GordenPPTSkill
A Python-based PPT builder that provides 17 pre-designed Chinese business...
zarazhangrui/beautiful-html-templates
A curated library of 34 reusable HTML slide templates designed specifically for...
nicobailon/visual-explainer
visual-explainer is an agent skill that converts terminal output into styled,...
| Repository | Stars | Week Δ | Language | Score | Updated |
|---|---|---|---|---|---|
|
|
7k | +253 | HTML | 8/10 | 3mo ago |
|
|
20.8k | — | HTML | 8/10 | 1mo ago |
|
|
2.6k | — | Python | 7/10 | 3w ago |
|
|
3.6k | — | HTML | 7/10 | 1mo ago |
|
|
9.2k | — | HTML | 7/10 | 2w ago |
|
|
25k | — | JavaScript | 8/10 | 2w ago |
Higher star count suggests broader adoption. Likely targets a wider audience beyond AI agents. html-ppt-skill differentiates on AgentSkill packaging and deeper theme variety.
Direct category competitor, also an AgentSkill for HTML presentations with significantly more stars. html-ppt-skill appears to offer more themes and layouts on paper but trails on adoption signals.
Different focus (explanatory visuals vs. slide decks), but overlapping use case in AI-generated content. Not a direct substitute.
Python-based AgentSkill for similar purpose. html-ppt-skill's pure static HTML approach may offer simpler agent integration for output rendering.
Established HTML presentation framework with large ecosystem. html-ppt-skill is not a direct competitor — it layers opinionated templates on top of the static-file approach rather than providing a programmable framework.






