DOCSEARCHbyAlgolia
Submit Docs

Skills Collection for Claude Code

repository·main·Indexed 42 days ago

https://github.com/wwwzhouhui/skills_collection

A collection of specialized 'Skills' (Skill.md files) for Claude Code v2.0+ to extend agent capabilities in automation, content creation, video production, and data scraping. It includes the AI Teaching Media skill suite, which provides six interconnected capabilities: ai-image-generator, tech-article-diagram, edu-subject-infographic, edu-teaching-animation, short-video-cover, and article-explainer-video.

Tokens
273K
Snippets
506
Records
1K
Agent score
55%

What's inside Skills Collection

  1. Overview of jimeng-mcp-server tools

    main

    The jimeng-mcp-server provides four primary tools for AI-driven visual content generation. All tools return a JSON response containing the generated content URL(s).

    1. text_to_image: Generates high-quality images from text descriptions using the Jimeng 4.5 engine.
    2. image_composition: Merges and blends multiple images (2-5 images) using intelligent fusion.
    3. text_to_video: Creates short videos (typically 3-5 seconds) from text prompts.
    4. image_to_video: Adds motion and animation effects to static images.

  2. Overview of the AI Teaching Media Integrated Skillset

    main

    The ai-teaching-media skill is an integrated package designed to transform content (knowledge points or technical articles) into visual teaching assets. It coordinates 6 specialized sub-skills to handle the workflow from content understanding to image generation, infographics, animations, and final video production.

    Sub-skill Capabilities

    Sub-skillPurposeTypical InputOutput
    ai-image-generatorBase image generation layer (called by others)prompt / reference imagePNG
    tech-article-diagramIntelligent diagrams for technical articlesMarkdown long-form text16:9 illustrations
    edu-subject-infographicSubject-specific infographicsKnowledge point name9:16 infographic
    edu-teaching-animationTeaching animations + voiceover videosSubject conceptSilent MP4 / Voiceover MP4
    short-video-coverVertical covers for short videosSpoken script/copy3:4 cover image
    article-explainer-videoChapter-based explainer videosTechnical article1080p MP4

    Execution Principles

    • Route first, execute later: Always read the specific SKILL.md of the target sub-skill before execution.
    • Thematic Consistency: Maintain consistent color schemes and terminology across all assets for the same topic.
    • Unified Image Generation: All sub-skills should use ai-image-generator for actual image creation rather than implementing their own APIs.
    • Dependency Awareness: If required API keys or tools (like ffmpeg) are missing, inform the user immediately instead of failing silently.

  3. Overview of NoBase Integration

    main

    NoBase is a backend API integration platform by XunAn Technology. It allows users to integrate various functional systems via API without writing backend code. It supports:

    • Application and User systems (Login, Registration, etc.)
    • Card/Redemption code, Email, and Forum systems
    • AI Large Model dialogue interfaces and Role-play (Prompt Agent) systems
    • Cloud Document systems (Text, JSON, Webpages)

  5. Overview of xiaohuihui-tech-article generator

    main

    The xiaohuihui-tech-article is a specialized generator designed to create technical tutorial articles following the specific writing style and structural standards of the 'Xiaohuihui' (小灰灰) WeChat Official Account.

    Key capabilities include:

    • Standardized Structure: Automatically generates a complete article including an author business card, preface, project introduction, deployment practice, summary, and past article recommendations.
    • Technical Content Generation: Produces detailed step-by-step guides (Environment Prep → Dependency Installation → Configuration → Implementation → Testing) and deep summaries (300-500 words in a single paragraph).
    • Automated Visuals: Uses the Gemini-3-Pro-Image-Preview model via two channels (Self-built API or the Gemai public station api.gemai.cc) to generate images.
    • Image Hosting: Automatically uploads generated images to Tencent Cloud COS (Cloud Object Storage).
    • Tone & Style: Employs a conversational, friendly technical tone (using phrases like "呵呵", "好家伙", "手把手教").

  6. Overview of voice-to-video pipeline

    main

    The voice-to-video skill package converts a spoken script into a fully synchronized MP4 video. The process follows a deterministic pipeline where audio acts as the master clock, driving HTML-based animations and subtitles.

    Workflow:

    1. TTS Generation: scripts/tts.py converts text to speech (using edge-tts) and produces a timeline.json containing word-level timestamps.
    2. Composition: A user writes a composition.html file using the provided assets/engine.js to define visual scenes that align with the audio timestamps.
    3. Rendering: scripts/render.py uses a headless Chromium browser (via Playwright) to seek through the timeline and capture deterministic frames.
    4. Encoding: ffmpeg encodes the frames and audio into a final .mp4 video.

  7. Overview of Juggle Microservice Orchestration Framework

    main

    What is Juggle?

    Juggle is a low-code/no-code and AI-driven microservice orchestration and system integration platform. It is designed to solve pain points such as microservice interface orchestration, third-party system integration, standard product customization, and data cleaning.

    Key Features

    • Zero-code Workflow Construction: Easy to get started with visual orchestration.
    • Private Deployment: Supports full private deployment to ensure data security.
    • High Performance: Optimized for high concurrency.
    • Protocol Support: Orchestrates interfaces using HTTP, Dubbo, WebService, etc.
    • Scripting & Data: Supports enhancement via Groovy, JavaScript, Python, and Java, and connects to various data sources like MySQL and Dameng.

    Getting Started

    To use Juggle, download the latest version from the Juggle Official Website and follow the Juggle Installation Manual.

  8. Overview of video-agent-kit

    main

    video-agent-kit is an automated video editing toolkit designed to provide specialized capabilities to AI agents (like Claude Code) via MCP (Model Context Protocol) tools and a CLI. It enables media probing, cloud-based speech transcription and synthesis, video frame observation, high-FPS segment review, basic video operations, timeline validation, preview rendering, and Quality Control (QC).

    Design Principles

    • Visual Observation: The toolkit does not assume the main model can receive video files directly. Instead, it extracts frames and creates a timestamped contact sheet, which is returned to the agent as an image result.
    • Self-Contained Logic: Frame extraction and indexing logic is built-in (mcp/ve_tools/video_observe.py). It does not construct image_url requests or call external inference services internally.
    • Scope: It processes user-provided video/audio assets. It does not perform asset library retrieval or image/video generation.
    • Automation: In automated data generation scenarios, the toolkit does not wait for user clarification or confirmation.

  9. Overview of Cherry Studio

    main

    Cherry Studio is a multi-model desktop client designed to improve productivity across various scenarios. It supports Windows, macOS, and Linux and integrates mainstream LLM cloud services, AI Web services, and local model execution.

    Key features include:

    • Full Markdown rendering
    • Agent creation
    • Translation capabilities
    • File uploads and multi-modal conversations
    • Flexible theme options and a user-friendly interface.

  10. Overview of ComfyUI Prompt Assistant

    main

    The ComfyUI Prompt Assistant is an extension designed to enhance prompt engineering within ComfyUI using LLM APIs (like SiliconFlow). It provides the following core features:

    • Prompt Translation: One-click bidirectional translation between Chinese and English without manual language settings.
    • Prompt Expansion: Automatically expands and polishes prompts based on specific rules for different image or video generation models.
    • Image-to-Prompt (Interrogation): Uses vision model APIs to convert image content into text prompts.
    • Tag Presets: Allows saving and quickly inserting common prompt tags or trigger words.

  11. Overview of Tech Article Diagram Generator

    main

    The Tech Article Diagram Generator is a tool designed to automatically add 4-5 conceptual diagrams (16:9 aspect ratio) to long-form technical articles (3000-8000 words). It identifies key locations in your text—such as abstract concepts, workflows, comparisons, or architectures—and generates visual aids in one of four distinct styles.

    Supported Styles

    • notebook: Hand-drawn notebook style. Black ink, green positive accents, red emphasis, on grid paper.
    • infographic: Professional infographic style. Off-white background, dark brownish-red titles, blue/orange dual-tone.
    • executive-tech: Modern high-end tech/business style. Off-white background, deep indigo primary color, card-based UI, dashboard charts, and magazine-style typography.
    • cozy-handdrawn: Warm hand-drawn card style. Off-white paper texture, black hand-drawn lines, pastel rounded cards, and light manga-style layouts.

    Note: If you do not specify a style, Claude will prompt you to choose one of these four before proceeding.

  12. Overview of SiliconFlow API Skills

    main
    The siliconflow-api-skills skill provides comprehensive assistance for developing with the SiliconFlow cloud service platform. It covers Large Language Model (LLM) API calls, image generation, vector models, and specific integration patterns like using SiliconFlow within Claude Code. It supports various API modes including Chat Completions and Streaming.