last30days-skill for Claude Code and Codex: research 30 days of public reactions

last30days-skill is a single skill that gathers reactions to one topic across several platforms over the last 30 days and turns them into a summary with sources. This post covers the problem it solves, who it suits, how to install it and start a first run, and the one thing to check before you rely on it.

What problem it solves

When a product launches or a competitor ships something, the reactions are scattered. Some sit in Reddit threads, some in X posts, some in YouTube comments and Hacker News discussions, and the same story often shows up in all of them under different titles. Reading each place by hand takes hours, and the result is usually a list of links rather than an answer to how people actually reacted.

The skill does the gathering and the grouping. It searches Reddit, X, YouTube, Hacker News, GitHub, and the web for recent reactions, merges discussions that cover the same story across sources, and ranks them by engagement, relevance, and recency. The summary it writes attaches sources to each claim, so you can cite it in a report instead of rewriting it from memory.

Who it fits

The chapter rates four situations. Three are enough to decide.

Situation	Fit	Reason
Researching launch reactions	Good fit	Collects comments and reaction counts
Comparing competing tools	Good fit	Supports comparison requests
Analyzing internal sales	Poor fit	Focuses on public reactions

Watching weekly trends sits in the middle. It works, but it needs extra setup to run on a schedule.

Install and first run

Install it through the Claude Code marketplace, or register it globally for Codex.

# Install through the Claude Code marketplace
/plugin marketplace add mvanhorn/last30days-skill
/plugin install last30days
# Install globally for Codex
npx skills add mvanhorn/last30days-skill -g -a codex

After installation, type your research topic after the command.

/last30days [topic to research]

The run first finds sources, then groups and ranks the reactions, then writes a summary you can cite. Reddit, Hacker News, Polymarket, and GitHub work without extra setup, so a first run on a public topic needs nothing else. If you want a file to share, the summary can be exported as a single-file HTML briefing or as versioned JSON. Before a run that touches cookies or files, --preflight shows the configuration and the planned actions without doing anything.

What trips people up

The catch is that the sources do not all cost the same. Some require a ScrapeCreators key with usage-based billing after the first 10,000 calls, X search can use your browser login cookies, and the YouTube source needs a separate brew install yt-dlp. The skill also runs local scripts that require Python 3.12 or later. One more setting deserves attention before you build on it: the README says the library’s --publish option defaults to public publishing, which creates an HTML index and briefing pages meant for static hosting such as GitHub Pages. Check that flag before you run research on anything you would not want published.

Where this comes from

This post adapts part of chapter 03-01 of The Claude Code & Codex Skills Guidebook. The full chapter covers the options for hiring signals and topic discovery, and how to store results in SQLite for scheduled watchlists and daily or weekly digests, and the book has 52 more skills like this one. Read the full chapter in The Claude Code & Codex Skills Guidebook on Gumroad.