Machine Learning

Create or edit Django templates with cotton components, template partials, and proper Tailwind usage. Use when creating/editing HTML templates, adding components, or when the user mentions templates, UI, or frontend work.

Generate UML diagrams including class, sequence, and use case diagrams. Use when modeling software systems or documenting architecture.

Write performance-friendly HTML pages. Use when optimizing page load, adding resources, configuring caching, or improving Core Web Vitals scores.

Guidance for building fullstack apps with Vite (React + TypeScript) frontend and FastAPI backend. Use when demos need a web UI beyond what Streamlit provides.

Creating new StickerNest widgets from scratch. Use when the user asks to create a widget, build a widget, add a new widget, make a widget component, or implement widget functionality. Covers manifest creation, HTML templates, TypeScript module structure, Protocol v3.0 communication, and widget registration.

Validate and lint structured data files (YAML, JSON, TOML) for consistency, required fields, syntax errors, and schema compliance.

Runs open-source ML models via Replicate API for image generation, LLMs, and audio. Use when calling Stable Diffusion, Llama, Whisper, or other models without infrastructure management.

Frontend design assistant for users with zero/low-code background. This skill should be used when users request creating web interfaces, landing pages, dashboards, or any frontend UI; when users ask about frontend terminology, design elements, or architecture concepts; when users want to understand CSS, layout, typography, or color choices. Emphasizes distinctive aesthetics over generic AI-generated designs, teaches concepts in context, and adapts to user skill level (pure HTML/CSS or React/Vue).

Generate PR documentation with UML diagrams based on Git branch differences. Explains modification reasons, content, and implementation. Uses Mermaid to draw architecture, class, sequence diagrams with exact code element names matching actual code.

单细胞分析最佳实践集合 - 100%完整覆盖（410个文件：407个HTML文档+3个项目文档）

Use AI to recreate PDF page as semantic HTML. Consumes three inputs (PNG image, parsed text, ASCII preview) for complete contextual understanding and accurate generation.

Generates index.html as the application entry point. Includes meta tags, favicon links, and root div for Vue app mounting.

This skill should be used when the user asks "Chart.js tooltips", "Chart.js tooltip callbacks", "custom Chart.js tooltip", "HTML tooltip Chart.js", "Chart.js tooltip formatting", "Chart.js tooltip position", "Chart.js tooltip label", "external tooltip", "Chart.js tooltip styling", or needs help configuring tooltips in Chart.js v4.5.1.

Playwright end-to-end testing patterns including page object models, test scenarios, visual regression, and CI/CD integration. Use when building E2E tests, testing web applications, automating browser interactions, implementing page objects, running Playwright tests, debugging E2E failures, or when user mentions Playwright, E2E, browser automation, page object model, POM, visual regression, or end-to-end testing.

Analyze and update Python dependencies in pyproject.toml, checking for compatibility and security vulnerabilities. Use when: updating dependencies, checking security issues, dependency analysis, version pinning, pip-audit, outdated packages.

Generate HTML research reports with embedded data, charts, and analysis.Use when creating final research deliverables. Supports single comprehensivereports or multiple focused reports. Handles styling, structure, and outputto reports/ directory.

Enforce mobile-first responsive design for all frontend changes. Covers TailwindCSS, vanilla CSS, container queries, dvh units, and framework-specific patterns. Trigger on any HTML, CSS, template, or UI component modifications.

Convert audio files to text using parakeet-mlx, NVIDIA's Parakeet automatic speech recognition model optimized for Apple's MLX framework. Run via uvx for on-device speech-to-text processing with high-quality timestamped transcriptions. Ideal for podcasts, interviews, meetings, and other audio content. This skill is triggered when the user says things like "transcribe this audio", "convert audio to text", "transcribe this podcast", "get text from this recording", "speech to text", or "transcribe this wav/mp3/m4a file".

Create properly formatted Claude Code skills with YAML frontmatter, directory structure, and best practices. Use when creating new skills, converting documentation to skills, or learning how to write skills.

스킬 건강도 진단 및 유지보수 가이드. 500줄 규칙, skill-rules.json 동기화, YAML 검증을 수행합니다.