A powerful research skill that scrapes the last 30 days of social media and web data to turn you into an instant expert on any topic.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install last30days-skill
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install last30days-skill using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is last30days?
The last30days skill is a sophisticated research engine designed for AI agents. It goes beyond standard web searches by tapping into the live pulse of the internet—specifically Reddit, X (Twitter), and YouTube. By analyzing engagement metrics like upvotes, likes, and views, it identifies what real people are actually discussing right now. This is a premier example of how Openclaw Skills can transform a standard LLM into a specialized research assistant capable of synthesizing social sentiment and technical trends into actionable knowledge.
Whether you are looking for the latest AI video tools or tracking a breaking news story, this skill provides a structured workflow to parse user intent and deliver high-signal data. It effectively bridges the gap between static training data and the rapidly evolving digital landscape, making it an essential tool for developers and researchers using Openclaw Skills.
last30days Use Cases
- Prompt Engineering: Discover the latest techniques and copy-paste prompts for tools like Midjourney, ChatGPT, or Nano Banana Pro.
- Trend Analysis: Identify trending products, libraries, or news across Reddit and X before they hit mainstream blogs.
- Product Recommendations: Get a ranked list of tools or services based on actual user mentions and engagement signals.
- Market Research: Understand public sentiment and debates surrounding a specific brand, celebrity, or technology.
How last30days Works
- Intent Parsing: The agent extracts the topic, target tool, and query type (News, Recommendations, or Prompts) from the user input.
- Multi-Platform Scraping: A Python script executes in the foreground to gather data from Reddit, X, and YouTube, including video transcripts.
- Web Enrichment: The skill performs a targeted WebSearch for blogs, tutorials, and documentation to supplement social data.
- Judge Agent Synthesis: An internal process weights social engagement (likes/upvotes) higher than generic web content to find the strongest signals.
- Expert Delivery: The agent presents a stats-backed summary and enters an Expert Mode to answer follow-up questions or generate tailored prompts.
last30days Setup
To get started with this skill in your Openclaw Skills environment, ensure you have an OPENAI_API_KEY configured. You will also need python3 and node installed. If you want video transcript analysis, ensure yt-dlp is available in your path.
# Run the research script manually if needed
python3 scripts/last30days.py "your topic" --emit=compact
last30days Data Schema & Taxonomy
The skill organizes research data across three primary social dimensions and a supplementary web layer. It uses a custom scoring system based on platform-specific engagement.
| Source | Data Captured | Metadata Captured |
|---|---|---|
| Thread Title | Upvotes, Comment Count, Subreddit | |
| X (Twitter) | Post Text | Likes, Reposts, @handle |
| YouTube | Video Title | Views, Likes, Transcript Snippets |
| Web | Article/Blog | Publication Name, URL |
last30days Advanced Features
- Engagement Weighting: Automatically prioritizes sources with high social validation (likes and upvotes) to filter out noise.
- Transcript Analysis: Deep-dives into YouTube video content using transcript extraction for technical precision.
- Custom Lookback: Support for
--days=Nto narrow research to a weekly roundup or broaden it to a full month. - Depth Control: Choose between
--quickfor rapid summaries or--deepfor comprehensive multi-hundred-source analysis. - Expert Mode Persistence: Once research is complete, the agent maintains its expert context for the rest of the session, allowing for deep follow-ups without re-searching.
SKILL.md
Loading
Related Openclaw Skills
Markdown Browser
A specialized orchestration layer that processes web fetch results into privacy-compliant, normalized markdown for AI agents.
Swarmind
Swarmind is a collaborative task management skill that enables AI agents and humans to work together through a structured Kanban workflow and secure API-driven orchestration.
Swarm Kanban
A multi-agent collaborative task management system featuring a Kanban workflow for seamless agent-to-agent and human-agent interaction.
Xiaohongshu Video Publish
An automated browser skill to streamline video publishing on the Xiaohongshu creator platform.
Claw Conductor
An always-on autonomous development orchestrator that triages requests, decomposes complex tasks, and manages multi-model parallel execution.
comfy-cli
A comprehensive command-line interface for installing, configuring, and scaling ComfyUI instances and their ecosystem of nodes and models.
Featured*
