summarize - Claude Skill Details | SkillHub

# summarize - Claude Skill Details | SkillHub [Main Site ↗](https://www.skillhub.club/web) [Run any skills on Happycapy.ai](https://happycapy.ai/?via=skillhub)[Get Started →](https://happycapy.ai/?via=skillhub) [ SKILLHUB](https://www.skillhub.club/)[All Skills](https://www.skillhub.club/skills)[🚀Stacks](https://www.skillhub.club/skill-stacks)[👑KOL](https://www.skillhub.club/skills/kol)[🔥Hot](https://www.skillhub.club/hot)[🦞OpenClaw](https://www.skillhub.club/web/openclaw) Products [Launch App](https://www.skillhub.club/app/skills)Login English [Back to List](https://www.skillhub.club/skills)[From collection "openclaw"](https://www.skillhub.club/skills/openclaw-openclaw) # summarize by openclaw 342.7k 15.6k[GitHub](https://github.com/openclaw/openclaw#skills-summarize) Summarize or extract text/transcripts from URLs, podcasts, and local files (great fallback for “transcribe this YouTube/video”). ### User Request input Node User provides input (URL, file path, or YouTube link) with optional flags ### Parse Input & Flags processing Node Extract source type, determine if --extract-only, --youtube, --json, --length flags are set ### Source Type? decision Node Determine whether input is URL, local file, or YouTube link ### Load Config config Node Read ~/.summarize/config.json for default model and settings ### API Keys Present? decision Node Check for required API keys based on selected model provider ### Process Local File processing Node Read and extract text from PDF, TXT, DOCX, etc. using local parsing ### Fetch URL Content processing Node Download and extract main content from webpage using built-in extractor ### Firecrawl Needed? decision Node Evaluate if site is blocked/JS-heavy requiring Firecrawl fallback ### Use Firecrawl processing Node Call Firecrawl API with FIRECRAWL_API_KEY for complex sites ### YouTube Flag Set? decision Node Check if --youtube auto flag is present for YouTube URLs ### Extract Only? decision Node Check if --extract-only flag is set for transcript extraction ### Use Apify Fallback processing Node Use Apify API with APIFY_API_TOKEN for YouTube transcript extraction ### Built-in YouTube processing Node Use built-in YouTube extraction without external APIs ### Select AI Model config Node Choose model based on config, flags, or default (gemini-3-flash-preview) ### Call LLM API processing Node Send extracted content to selected AI model with length/token constraints ### Format Output processing Node Apply --json flag formatting or pretty markdown based on flags ### Return Summary output Node Deliver final summary/transcript to user ### Return Transcript output Node Deliver extracted transcript (no summarization) ### API Error output Node Return error message about missing API keys [React Flow](https://reactflow.dev/) Press enter or space to select a node.You can then use the arrow keys to move the node around. Press delete to remove it and escape to cancel. Press enter or space to select an edge. You can then press delete to remove it or escape to cancel. Live Logic Circuit ### Output Preview Copy Output output_preview.md ``` # Summary: "The Future of AI Agents" (YouTube Video) ## 📊 Key Insights - **Agentic workflows** are shifting from single-prompt interactions to multi-step reasoning with tool use - **Three emerging patterns**: Reflection (self-critique), Tool use (API calls), and Multi-agent collaboration - **Critical challenge**: Maintaining consistency across long reasoning chains ("hallucination drift") ## 🎯 Core Argument The next breakthrough won't be larger models, but better **orchestration systems** that can: 1. Decompose complex problems into sub-tasks 2. Validate intermediate results 3. Recover from errors autonomously 4. Integrate specialized tools (calculators, code executors, web search) ## 🔧 Technical Implementation ```python # Example agentic loop pattern for attempt in range(max_retries): plan = llm.generate_plan(task) for step in plan: result = execute_tool(step) validation = llm.validate(result, step) if not validation.passed: llm.repair(step, result) # Self-correction ``` ## ⚠️ Limitations & Risks * **Cost scaling**: Each reflection/tool-use cycle adds latency and API costs * **Over-reliance**: Agents may get stuck in infinite correction loops * **Security**: Tool access requires careful sandboxing ## 🚀 Practical Recommendation Start with **single-agent, single-tool** workflows before scaling to multi-agent systems. The `summarize` tool itself demonstrates effective tool use by combining content extraction with LLM summarization. _Duration: 24:18 | Transcript available with `--extract-only`_ 10/10 Security Low security risk, safe to use - Clarity - Practicality - Quality - Maintainability - Innovation Development Compatible Agents  Claude Code ~/.claude/skills/  Codex CLI ~/.codex/skills/  Gemini CLI ~/.gemini/skills/ O OpenCode ~/.opencode/skills/ O OpenClaw ~/.openclaw/skills/  GitHub Copilot ~/.copilot/skills/  Cursor ~/.cursor/skills/ W Windsurf ~/.codeium/windsurf/skills/ C Cline ~/.cline/skills/ R Roo Code ~/.roo/skills/ K Kiro ~/.kiro/skills/ J Junie ~/.junie/skills/ A Augment Code ~/.augment/skills/ W Warp ~/.warp/skills/ G Goose ~/.config/goose/skills/ SKILL.md Evaluation Usage Guide SKILL.md # Summarize Fast CLI to summarize URLs, local files, and YouTube links. ## When to use (trigger phrases) Use this skill immediately when the user asks any of: * “use summarize.sh” * “what’s this link/video about?” * “summarize this URL/article” * “transcribe this YouTube/video” (best-effort transcript extraction; no `yt-dlp` needed) ## Quick start ```bash summarize "https://example.com" --model google/gemini-3-flash-preview summarize "/path/to/file.pdf" --model google/gemini-3-flash-preview summarize "https://youtu.be/dQw4w9WgXcQ" --youtube auto ``` ## YouTube: summary vs transcript Best-effort transcript (URLs only): ```bash summarize "https://youtu.be/dQw4w9WgXcQ" --youtube auto --extract-only ``` If the user asked for a transcript but it’s huge, return a tight summary first, then ask which section/time range to expand. ## Model + keys Set the API key for your chosen provider: * OpenAI: `OPENAI_API_KEY` * Anthropic: `ANTHROPIC_API_KEY` * xAI: `XAI_API_KEY` * Google: `GEMINI_API_KEY` (aliases: `GOOGLE_GENERATIVE_AI_API_KEY`, `GOOGLE_API_KEY`) Default model is `google/gemini-3-flash-preview` if none is set. ## Useful flags * `--length short|medium|long|xl|xxl|<chars>` * `--max-output-tokens <count>` * `--extract-only` (URLs only) * `--json` (machine readable) * `--firecrawl auto|off|always` (fallback extraction) * `--youtube auto` (Apify fallback if `APIFY_API_TOKEN` set) ## Config Optional config file: `~/.summarize/config.json` ```json { "model": "openai/gpt-5.2" } ``` Optional services: * `FIRECRAWL_API_KEY` for blocked sites * `APIFY_API_TOKEN` for YouTube fallback Source: [https://github.com/openclaw/openclaw#skills-summarize](https://github.com/openclaw/openclaw#skills-summarize) Content curated from original sources, copyright belongs to authors Grade B -AI Score Best Practices Best Practices:A ONE-CLICK Copy Export One-Click Install[Try this Skill](https://happycapy.ai/?via=skillhub) [0](https://www.skillhub.club/login?returnTo=%2Fskills%2Fopenclaw-openclaw-summarize "Login to favorite") User Rating [Login](https://www.skillhub.club/login?returnTo=%2Fskills%2Fopenclaw-openclaw-summarize)to rate this Skill USER RATING 0 UP 0 DOWN File Explorer 1 SKILL.md WORKS WITH  Claude  Codex  Gemini O OpenCode O OpenClaw  Copilot  Cursor W Windsurf C Cline R Roo K Kiro J Junie A Augment W Warp G Goose [ SKILLHUB](https://www.skillhub.club/) Discover, browse, and one-click copy the best Agent Skills. 15,000+ curated skills, LLM graded for efficiency. Stay Updated ### Quick Links * [All Skills](https://www.skillhub.club/skills) * [Rankings](https://www.skillhub.club/rankings) * [API Docs NEW](https://www.skillhub.club/docs/api) ### About * [GitHub](https://github.com/keyuyuan/skillhub-awesome-skills) * [Claude Docs](https://docs.anthropic.com/) * [Terms & Conditions](https://www.skillhub.club/terms-and-conditions) ### Contact * [Report Bug](mailto:feedback@skillhub.club?subject=Bug%20Report%20-%20SkillHub) * [Contact Us](mailto:contact@skillhub.club?subject=Contact%20-%20SkillHub) Content sourced from GitHub, copyright belongs to original authors © 2026 SkillHub. All rights reserved. 👋 ## Welcome to SkillHub! Looks like this is your first time here. Want a quick 5-minute intro to Claude Skills? You’ll immediately understand what they are and why they matter. Start Learning Skip for now - [x] Don't show this again ## SkillHub Desktop is Here! Brand New Desktop App We've launched a brand new desktop app for a better SkillHub experience: * •One-click install Skills to Claude Code * •Manage your Skill collections locally * •Faster response time * •Supports macOS / Windows / Linux Download Now Maybe Later [Run any skill with one click](https://happycapy.ai/?via=skillhub)