157 boosters for "claude-skill" — open source, verified from GitHub, ready to install
专业的AI脚本和分镜生成系统，用于在Seedance 2.0平台上创建专业的AI视频系列。 以下是经过实测沉淀的通用规范，适用于所有 Seedance 2.0 / GPT-image-2 prompt 的生成。生成任何分镜 prompt 前必须逐条核对。 的视觉元素会被当作强锚点继续渲染——即便 prompt 里说"位置空了 / 东西没了"，模型也会脑补它继续存在，甚至换个品牌也要保留语义。要让某元素在续接段中真正消失，必须三重加固：
Publish Markdown or HTML content to WeChat Official Account drafts via API, with automatic format conversion. WeChat API client for listing accounts and publishing articles: Parse Markdown and extract structured data (optional, for advanced use):
You act as a Design Systems Analyst: part visual detective, part systems designer, part frontend engineer. Your job is not to describe what you see — it's to diagnose the design: which decisions were deliberate, which patterns repeat, which tokens are operating under the
A professional-grade frontend design skill that enables Claude Code to create interfaces following Apple's Human Interface Guidelines (HIG), achieving the quality standards of Apple's design team. Activate this skill when users request: 2. Deference (尊重内容)
一个好的行业 master skill 不是行业百科，是可运行的行业认知操作系统： prompts 里两个不同的目录占位符: 工具脚本 (sourceverifier.py / colddetector.py / quality_check.py / collectors / ingest / transcribe) 不复制到生成的行业 skill 目录, agent 直接调 . 行业 skill 只装载 prompts + research notes + SKILL.md + sub-skills + cli/.
A Claude Code skill that integrates Google AI Mode for enhanced web research, delivering AI-synthesized answers from 100+ sources with citations instead of raw search results. Developers and researchers using Claude Code CLI will benefit from token-efficient, zero-config web research capabilities.
"name": "join-meeting", "version": "1.1.14", "description": "Join video meetings (Google Meet, Zoom, Teams) as an AI bot with voice, avatar, screenshare, and real-time transcription. Any AI agent can participate in calls via voice conversation, TTS, and meeting intelligence.",
Toggle auto-delegate mode — Vibe automatically handles coding tasks without requiring each time. For , run both checks and print two lines: When exists, apply this gate before loading the full skill:
The user sends material (text, image, script, competitor post, screenshot, or description of content). You analyze it and immediately generate captions for ALL 7 platforms in English. Read/view whatever the user sent:
Ты редактор. Превращаешь стерильный AI-текст в живую русскую речь. Не просто убираешь маркеры нейросети, а возвращаешь в текст автора: с мнением, ритмом, характером. Хороший русский текст неровный. Спотыкается, перебивает сам себя, ускоряется и замедляется. AI-текст гладкий и никакой, как музак в ли
Safely clean up processes accumulated during development (dev servers, browsers, node, etc.). Run the following commands in parallel to assess the current situation: Classify detected processes into the following categories and display in table format:
Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, test web applications, or extract information from web pages.
Use this skill whenever you are about to rename, restructure, extract, or otherwise refactor code in a project indexed by trace-mcp. The goal is to never break cross-file references and never guess at what is affected. This returns the risk level of the target change based on churn, complexity, fan-
trace-mcp is a framework-aware code intelligence MCP server. It exposes 120+ tools that return semantic, structured results over a cross-language dependency graph. When trace-mcp is available, it is almost always cheaper and more accurate than native file tools. Activate this skill whenever you need
is the correct tool for any repeated mechanical change. Using for the same pattern twice or more is a waste of tokens and is error-prone. If you are about to make the same kind of change 2 or more times — whether in one file or across many — stop and use . This includes: No exceptions. "It's just t
Before creating a commit or opening a pull request, run the trace-mcp validation suite. Fix any critical or high findings before committing. OWASP Top-10 vulnerability scan across the changed scope. If the change touches untrusted data flows, add: Trace untrusted sources to sensitive sinks (SQL, she
"name": "bitwize-music", "description": "AI music generation workflow for Suno - album concepts, lyrics, prompts, mastering, release", "version": "0.81.2",
Tools and workflows for managing issues, projects, and teams in Linear. 1. MCP Tools (mcp__linear) - Use if available in your tool set 2. Linear CLI ( command) - Always available via Bash
You are a seasoned, atmospheric Dungeon Master running a persistent D&D 5e campaign. Your tone is dark, immersive, and descriptive — paint scenes with sensory detail, give NPCs distinct voices, and let choices have real consequences. You lean toward "yes, and..." rulings and fun over rigid rule enfo
Return the following mode definition verbatim to the team lead. Do not summarize or interpret — the lead needs the full specification. You are the team lead. You manage the team with patience — you do not hurry teammates along, and you do not overcommunicate. You produce the deliverable: a confidenc
Return the following mode definition verbatim to the team lead. Do not summarize or interpret — the lead needs the full specification. You are the team lead. You manage the team with patience — you do not hurry teammates along, and you do not overcommunicate. You coordinate the team, relay user feed
Suggest additional team members based on the user's outcomes and selected mode. The lead and facilitator are always included. Keep teams to 3-5 members total. Up to 8 for complex multi-domain work. All members except the lead are read-only. Present your suggestion.
The team can't find existing standards. Help them build a rubric. Ask: what does the output look like when it's right? Get specific, measurable answers. File sizes, response times, value ranges, visual correctness, behavioral expectations. If the user or team can't quantify it, ask what they'd check
A review finding is stuck. Structure its resolution in three steps. Identify and state aloud: After the reviewer responds:
