Last30Days Skill
Stars: 35.4K ★ | Language: Python | License: MIT
AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web — then synthesizes a grounded summary.
Overview
A research skill for AI agents (Claude Code, etc.) that aggregates the “last 30 days” of any topic from multiple social and news sources, then produces a synthesized, sourced, grounded summary. Built for recency-aware deep research.
Data Sources
- Reddit — community discussion & sentiment
- X (Twitter) — real-time chatter & thought leaders
- YouTube — video content & creator perspectives
- Hacker News — tech/startup discourse
- Polymarket — prediction market odds & sentiment
- Web search — general news & articles
- Bluesky / Instagram / TikTok — additional social layers
Output
A structured report covering:
- What happened in the last 30 days on the topic
- Key voices, debates, and sentiment
- Prediction market signals (if applicable)
- Links to original sources for verification
Use Cases
- Trend spotting — what’s being discussed right now
- Market research — aggregate sentiment across platforms
- Competitive intelligence — real-time pulse on a space
- Investor research — grounded by Polymarket + news
Related
- Hermes Agent Skill Authoring — how to author similar skills
- Karpathy LLM Wiki — structured knowledge base approach
- Understand Anything — code-to-knowledge-graph