last30days

Research a topic the way a sharp human would skim the last month of the internet: pull what real people actually engaged with across Reddit, Hacker News, X/Twitter, YouTube, and GitHub within a recent time window, weight it by engagement and cross-source agreement, and synthesize a grounded brief with inline citations.

This is a lightweight, zero-dependency take on mvanhorn/last30days-skill. It runs entirely on the built-in WebSearch and WebFetch tools — no API keys, no Python engine.

When to use this vs. deep-research: reach for last30days when recency and community signal are the point (trend discovery, product/tool selection, "is X any good lately", sentiment, current events). Reach for deep-research when you need authoritative, fact-checked answers regardless of date.

Prerequisites

• WebSearch and WebFetch are built-in — no setup. If WebSearch isn't available in the session, load it first (ToolSearch → select:WebSearch,WebFetch).
• WebSearch is US-region and works best on indexed web. Reddit, Hacker News, GitHub, and YouTube are reliably searchable. X/Twitter is partially indexable — treat it as best-effort and lean on the others when it's thin.

Workflow

Step 0 — Scope the query