March 25, 2026Open SourceToolAgents

last30days-skill: AI Agent Research Skill Trending #1 on GitHub

A new open-source skill called last30days-skill has shot to #1 on GitHub Trending with over 1,300 stars per day. Built for Claude Code and OpenAI Codex, the skill gives AI agents the ability to research any topic across Reddit, X/Twitter, YouTube, TikTok, Instagram, Hacker News, Polymarket, Bluesky, and the broader web β€” then synthesize findings into a grounded narrative with real citations.

The skill works as a deep research engine covering the last 30 days across 10+ sources. When invoked, it finds what communities are upvoting, sharing, betting on, and discussing, then writes a cited report. It includes YouTube transcript extraction via yt-dlp, engagement data analysis, and a watchlist feature with scheduling β€” you can track competitors weekly with results accumulating in a local SQLite database.

What makes this notable for the agentic ecosystem is the pattern it demonstrates: agents gaining autonomous research capabilities across multiple real-time data sources. Rather than being limited to web search, the agent can tap into social signals, prediction markets, and video content to build a comprehensive understanding of any topic. The project is MIT-licensed and available at https://github.com/mvanhorn/last30days-skill.
← Previous
TurboQuant: Google's Compression Algorithm Cuts LLM Memory 6x with Zero Accuracy Loss
Next β†’
Check Point Launches AI Defense Plane to Secure the Agentic Enterprise
← Back to all articles

Comments

Loading...
>_