Frontend Slides
zarazhangrui/frontend-slidesSkill that builds animated, single-file HTML slide decks from scratch or from a PowerPoint, after showing you three visual style previews.
Recommended
Check first Depth 4/5
HTML Anything Claude skills
Local agentic HTML editor with 75 skills across 9 surfaces (magazine, deck, social cards) that drives a logged-in coding-agent CLI to write and preview HTML.
Our verdict Recommended with caveats
A clever local orchestrator with a genuinely large template library, but it runs your coding-agent CLI with its permission checks turned off by design.
What HTML Anything does
HTML Anything is a local web app, not a single skill: it detects which coding-agent CLIs are on your PATH (Claude Code, Codex, Cursor Agent, Gemini CLI, GitHub Copilot CLI, OpenCode, Qwen Coder, Aider, IBM Bob) and spawns one to write HTML for a chosen surface, streaming its output over SSE into a sandboxed iframe preview. The 75 bundled skill templates each describe one output type and style, for example article-magazine for Substack-style long-form layout, article-sketchnote-editorial for a six-panel visual narrative, blog-post, dashboard, pricing-page and several social-card formats for WeChat, X and Xiaohongshu. Templates are short (under 2 KB typically, with a few exceptions near 12 KB) and specify layout, typography and content ratios rather than full code. The app needs no API key because it reuses the login session you already have for your CLI of choice; it invokes several of those CLIs with flags that skip their normal permission prompts, such as Codex's --sandbox workspace-write, Gemini's --yolo and OpenCode's --dangerously-skip-permissions, since the orchestrating process is itself the privileged one. Output exports to WeChat, X, Zhihu, static HTML or PNG with one click.
Good for
- Converting markdown drafts into formatted HTML for specific publishing surfaces
- Users who already have a coding-agent CLI logged in and want a local UI around it
- Social card and magazine-style layouts for Chinese and English platforms
Not for
- Users without a supported coding-agent CLI already installed and authenticated
- Workflows that need the agent's normal permission prompts preserved per action
What's inside depth 3/5
81 skill folders averaging 1.7 KB, mostly short style templates with a few longer ones, plus a non-trivial detection and spawning app behind them; no tests or evals for the templates themselves.
- Skills
- 81
- Reference docs
- 32
- Scripts
- 0
- Tests / evals
- 0
- Templates
- 9
- Avg SKILL.md
- 1.6 KB
Counted from the repository file tree. How we score depth.
Dependencies and costs
- A supported coding-agent CLI already installed and logged in (Claude Code, Codex, Cursor Agent, Gemini CLI, Copilot CLI, OpenCode, Qwen Coder, Aider or IBM Bob)
- Node.js for the local app
Security review
Broad accessThis skill asks for more access than most, or sends data to another service. That is usually a design choice rather than bad intent, but you should know before you install it.
- Turns off permission prompts or safety checksThe app launches your installed coding-agent CLIs with permission-skipping flags (gemini --yolo, opencode --dangerously-skip-permissions, cursor-agent --force --trust, copilot --allow-all-tools, qwen --yolo, aider --yes-always) as documented in the supported-agents table.README.md:197
A local web app plus a library of about 75 HTML-generation skill templates (decks, dashboards, cards). By its README the app spawns your installed coding-agent CLIs with their permission prompts disabled and reuses your logged-in sessions, rendering output in a sandboxed iframe. We only have the README and template folders, not the app code, so what the app itself does beyond its description is unverified.
Worth knowing
- Relies on a CLI or package we could not reviewDepends on a locally installed, already logged-in coding-agent CLI (Claude Code, Codex, Cursor Agent, Gemini, Copilot, OpenCode, Qwen, Aider, Bob) and reuses its session; the app's own code (cli/, next/src) is not among the files we have.README.md:192
What we checked
The repository (379 files) on Oct 29, 2026. Not a full review: 254 of 379 files were not downloaded (large repository; we reviewed the skill folders); it relies on code outside the repository: The application code (cli/ and next/src TypeScript, API routes, agent spawning in next/src/lib/agents/argv.ts) is not in the files we have; only the README and the template skill folders were available; pnpm dependencies of the workspace (juice, modern-screenshot, papaparse, xlsx, etc.); Third-party agent CLIs that the app spawns.
NVIDIA SkillSpector static scan: 4 pattern hits (Agent Snooping 2, Supply Chain 1, Privilege Escalation 1). The curl-pipe-shell hit is sample text inside a docs-page example.html template (a fictional install snippet in a generated page), not an instruction to the agent. 'keyring' and 'skills/*/SKILL.md' hits are text inside example HTML decks.
A review, not a certification. Skills change after we check them — read the files before installing.
License
Apache-2.0 — commercial use allowed.
Can it run unattended?
Not really. It is an interactive local editor with a live preview meant for you to review before exporting.
Alternatives to HTML Anything
Answer me with HTML
QingYunA/answer-me-with-htmlTurns complex answers into a one-page HTML explainer: the model writes a short Markdown draft, and a bundled CLI generates layout, diagrams and dark mode.
Recommended
Check first Depth 3/5
md2wechat
geekjourneyx/md2wechat-skillA CLI plus agent skill that converts Markdown into WeChat Official Account HTML, makes covers and images, and pushes drafts to the WeChat draft box.
With caveats
Check first Depth 3/5
FAQ
What is HTML Anything?
HTML Anything is a local app that detects your installed coding-agent CLI and has it write HTML for a chosen surface, such as a magazine article, pricing page or social card, with a live sandboxed preview and one-click export.
Do I need an API key for HTML Anything?
No. It reuses the login session of whichever coding-agent CLI you already have authenticated (claude login, cursor login, gemini auth, and so on), so there is no separate API key to manage.
Is HTML Anything safe to run?
It runs your coding-agent CLI with permission-skipping flags and renders the generated HTML in a sandboxed iframe. That is a deliberate design trade-off for a local tool you control, but it means the CLI's own safety prompts are bypassed for that session.
HTML Anything vs nexu-io/open-design?
HTML Anything is described as the focused, agent-era editor from the same team; open-design is their larger, longer-running sibling project. Pick HTML Anything if the smaller, zero-API-key tool fits your workflow.
Sources: GitHub repository (stars, license, file tree, last push Oct 13, 2026). Review based on the README and SKILL.md files as published.