Analysis updated 2026-05-18
Research a person before a meeting by finding their recent tweets, podcast appearances, and GitHub activity.
Compare tools or products by seeing what real communities are saying and upvoting about them.
Understand breaking news events by aggregating what people are discussing across multiple platforms in real time.
Learn what a community actually thinks about a topic by searching engagement-ranked posts instead of SEO-optimized results.
| mvanhorn/last30days-skill | humansignal/labelimg | cookiecutter/cookiecutter | |
|---|---|---|---|
| Stars | 24,914 | 24,939 | 24,853 |
| Language | Python | Python | Python |
| Setup difficulty | hard | easy | easy |
| Complexity | 2/5 | 2/5 | 2/5 |
| Audience | vibe coder | data | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires API keys from 5+ services (Reddit, Twitter, YouTube, Polymarket) with varying approval timelines and rate limits, Claude Code integration adds another dependency layer.
Last30Days is an AI agent skill that acts as a search engine for what's actually happening on social platforms right now, not what editors decided was important, but what real people upvoted, liked, watched, and bet real money on. You install it as a plugin in Claude Code or compatible AI agent environments, then run a command like /last30days [topic] to get a synthesized briefing on any person, product, event, or topic from the past 30 days. It searches across Reddit (upvote-ranked comments), X/Twitter (engagement-ranked posts), YouTube (full transcript search), TikTok, Hacker News, GitHub, Polymarket (prediction markets where people bet real money), Bluesky, and more, all in parallel. The AI then ranks results by real engagement signals (upvotes, views, dollar-backed odds) rather than SEO, and synthesizes them into a brief. The key insight is that no single AI or search engine has access to all these platforms simultaneously, this skill bridges them by letting you bring your own API keys and browser sessions. People use it to research someone before a meeting (finding their recent tweets, podcast appearances, and GitHub activity rather than a stale LinkedIn), compare tools, understand breaking news events, or quickly learn what a community actually thinks about something. An optional HTML export feature lets you share the briefs. The skill installs with zero configuration for Reddit, Hacker News, Polymarket, and GitHub, a setup wizard unlocks the others. Written in Python.
An AI agent skill that searches Reddit, Twitter, YouTube, TikTok, and prediction markets to show you what people are actually talking about and betting on right now, not what algorithms decided to promote.
Mainly Python. The stack also includes Python, Claude Code, Reddit API.
Use freely for any purpose including commercial, as long as you keep the copyright notice.
Setup difficulty is rated hard, with roughly 1day+ to a first successful run.
Mainly vibe coder.
This repo across BitVibe Labs
Verify against the repo before relying on details.