> FREE STARTER (Whop): https://whop.com/agentdesk-1268/free-starter-last30days-trend-research
> FULL SYSTEM ($45): https://whop.com/agentdesk-1268/last30days-trend-research-cross-source-agent-skill — use code LAST30 for 30% off ($45 → $31.50).
## FREE vs FULL
| | FREE (this download) | FULL ($45) |
|---|---|---|
| Brief templates | 1 (AI tools) | 4 (crypto, AI tools, products, creators) |
| Source presets | 1 (free paths) | Full set, per niche |
| Ranking rubric | Summary in README | Full rubric |
| Doctor runbook | — | Yes |
| Daily-scan schedule | — | Yes |
| Prompt library | 3 prompts | Full library |
| Guides (EN + AR) | Quick README only | Yes |
## What this is
last30days-skill (mvanhorn) is an AI agent skill — a Python engine + SKILL.md — that researches any topic across Reddit, X, YouTube, TikTok, Hacker News, Polymarket, GitHub and the web, then synthesises a grounded summary ranked by what people actually did (upvotes, likes, engagement, real-money odds) instead of SEO. On 17 Nov 2026 it was #2 on GitHub Trending with 68,617 stars (v3.31.0, MIT licence). It also ships a "doctor" health check for broken sources.
This free lite edition gives you the first step: one brief template (AI tools), one source preset (free paths only), and the quickstart. The full pack adds 3 more brief templates, the full source presets, the ranking rubric, the doctor runbook, the daily-scan schedule, the full prompt library and the EN + AR guides.
## Free paths vs keyed paths
Free, no keys: Reddit · Hacker News · Polymarket · GitHub · arXiv.
Needs keys (optional): X · TikTok · Instagram — ScrapeCreators pay-as-you-go / X API. Start free; add a keyed source only when a question the free paths can't answer is worth paying for.
## Files you need (not included)
The upstream skill itself is NOT in this download — it is MIT open source, clone it from GitHub: https://github.com/mvanhorn/last30days-skill
No model files needed. No keys included — you add your own, if you choose to, in your own environment.
Built from public research; all content original. Star counts and trending ranks describe the upstream repo as seen on 17 Nov 2026 — they describe the market, not a promise.
