What does this skill do?

The AI Data Scraper Agent Skill creates an automated data collection agent that monitors any public source (job listings, prices, news, repositories), enriches the results with Gemini Flash, and stores them in Notion, Sheets, or Supabase. It learns from your decisions over time and runs for free on GitHub Actions without the need for servers.

Job Listings Monitor
Scrape LinkedIn, InfoJobs, and Indeed every 3 hours, filter results using AI based on your resume, and rate each job listing on a scale of 0 to 100 based on how well it matches your priorities.
Price Tracker
Monitor product prices on Amazon, MediaMarkt, or PcComponentes every 6 hours and receive alerts if they drop by more than 10%.
GitHub Repository Curator
Search the GitHub API for new repositories, summarize the README using AI, and rate their relevance based on your interests.
Sentiment-Based News Monitor
It scans media RSS feeds every hour, filters mentions of your company, and classifies sentiment as positive, neutral, or negative.

Usage examples

💼 Employment Monitor
Create an agent that monitors Python and machine learning job postings on LinkedIn every 3 hours and saves them to Notion with a match score based on my resume.
💰 Price Alert
Set up a scraper that tracks the prices of three products on Amazon and notifies me via Telegram if they drop by more than 10%.
🐙 Repos Curator
Create an agent that searches for new repositories on GitHub related to LLMs and agents, summarizes the README using AI, and saves them in Notion with stars and the last update date.
📰 Sentiment Analysis
Create a news monitor that checks RSS feeds from 10 media outlets every hour, filters out mentions of my company, and classifies the sentiment of each article.

Features

Three-tier architecture Collect → Enrich → Store Pattern with separate modules for scraping, AI, and the database.
AI with automatic fallback A Gemini Flash client that switches between models when the free quota runs out, with batch analysis of 5 items.
Feedback-Based Learning System Save your decisions (saved vs. rejected) in a JSON file and use them to bias future AI scores.
Automation with GitHub Actions Scheduled execution with cron—no servers, no costs. Automatically saves feedback with each run.
Multiple storage destinations Compatible with Notion, Google Sheets, and Supabase. Dedupes by URL to avoid duplicate entries.

Frequently asked questions

No. The agent runs on GitHub Actions using cron, which is free for public repositories. You don't need to pay for hosting or infrastructure.
Yes. Gemini Flash offers 500 requests per day and 1 million tokens per day in its free tier. In addition, the client automatically switches between models when the quota is exhausted.
Any public source: websites with static HTML (BeautifulSoup), sites with JavaScript (Playwright), paginated REST APIs, and RSS feeds. You just need to create one connector module per source.
The system saves the items you mark as "saved" or "rejected" in a JSON file. That history is used as a prompt that influences the AI's scores in future runs.
AI-Powered Data Scraper Agent

¿Prefieres escuchar el contenido? Genera la narración de audio con un clic.