Skill UI

This skill provides advanced techniques to significantly reduce API costs when using Firecrawl for web scraping. It covers best practices such as setting crawl limits, utilizing batch scraping for known URLs, implementing local caching, and monitoring credit usage. Learn how to efficiently scrape large sites while managing budgets and minimizing unexpected billing spikes.

Automate PandaDoc workflows from Claude Code so you can create documents from uploads or URLs, manage contacts and folders, configure templates and webhooks, and track document status without leaving the terminal.

Extract transcripts from YouTube videos, validate URLs and availability, then use STAR+R-I-S-E with youtube-transcript-api to craft comprehensive summaries and documentation for lectures, tutorials, and informational clips.

Compresses natural-language memory files such as CLAUDE.md, todo lists, and preferences into terse caveman-style text to reduce input token usage. Preserves code blocks, inline code, URLs, headings, and important technical terms while backing up the original to an out-of-tree data directory.

Command-line helper that pulls in URLs, local files, transcripts, podcasts, and YouTube links to deliver concise summaries or transcripts, with provider models and fetch fallbacks configurable for any workflow.

This skill aggregates and summarizes the latest news from multiple authoritative Chinese sources across various categories, including Politics, Finance, Society, World, Tech, Sports, and Entertainment. It utilizes web browsing tools to open specified URLs, take content snapshots, and extract key headlines and detailed summaries, providing the user with a concise, structured report.

This comprehensive skill is designed to retrieve content from various Feishu cloud documents, returning the raw Markdown format. It provides advanced logic for handling Wiki URLs by first determining the actual document type (Docx, Sheet, Bitable) before calling the appropriate specialized tool. Furthermore, it guides the user on how to separately download and process embedded media like images, files, and whiteboards using dedicated media tools.

This powerful skill analyzes various document types (PDF, Word, RTF) and web URLs. It generates a detailed report where every factual claim made in the analysis is accompanied by an inline screenshot from the source material, highlighting the exact cited text. This ensures users can visually verify the accuracy and source of every piece of information.

This skill summarizes the current conversation into a structured handoff document, ideal for passing work to a fresh AI agent or starting a new session. It avoids content duplication by referencing existing artifacts (PRDs, plans, issues, commits) using their paths or URLs, ensuring the receiving agent has a concise and actionable overview of the project's state, open decisions, and next steps.

A protocol for token-efficient, accurate, and structured web research. It guides the model to fetch minimum data when accessing URLs, GitHub repositories, or performing web searches, preventing token waste.

This advanced skill downloads Instagram posts, carousels, and Reels at their original, high resolution directly from the CDN. It supports batch processing of multiple URLs, automatically creates organized folders for each post, and stitches multi-slide carousels into a single PDF file, making content archiving and aggregation seamless.

This skill automates the NotebookLM AI research tool using browser automation via the Claude Chrome extension. It allows users to programmatically create new notebooks, add multiple sources (URLs, documents, text), and trigger complex output generations, such as mindmaps, audio overviews, and briefing documents, without manual clicking. Ideal for researchers and builders managing high volumes of source material.