# mvanhorn/last30days-skill

AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary

Repository: https://github.com/mvanhorn/last30days-skill
Canonical: https://ross.abutalabs.com/products/last30days-skill
Language: Python
License: MIT
License Family: permissive
Topics: ai-prompts, claude, claude-code, reddit, twitter, hackernews, polymarket, research, youtube, ai-skill, bluesky, clawhub, deep-research, instagram, openclaw, recency, social-media, tiktok, trends, web-search
Last push: 2026-08-26T22:09:45+00:00

## Health v2 (maintenance only)
Score: 82/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 98, longevity 15
- inputs: {"age_days": 222, "days_push": 7, "days_rel": 15, "gap_med": 0.5, "n_releases_24m": 41}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 59340, forks 5190 (observed 2026-08-28T04:12:18.732452+00:00)

## What it is
An AI agent skill/plugin that researches any topic across Reddit, X, YouTube, Hacker News, Polymarket, TikTok, and the web in parallel, scoring results by real user engagement. An AI agent judge then synthesizes the gathered signals into a single grounded, recency-focused brief.

## Use cases
- research what people are saying about a topic in the last 30 days
- get a briefing on a person before a meeting
- track trends across Reddit, X, and TikTok
- check Polymarket odds and community sentiment on events
- find the most upvoted discussions about a technology
- replace manual multi-platform searching with one agent command
- summarize recent community reactions to a product launch

## When to choose
- you use Claude Code, Codex, Cursor, or another agent-skills host and want one-command multi-platform research
- you need recency-weighted summaries grounded in real engagement signals rather than editorial search results
- you want to bridge walled-garden platforms (Reddit, X, YouTube, TikTok) that no single AI can search natively

## When to avoid
- you need a standalone application or hosted service rather than an agent skill
- you require official API compliance for all platforms, since it relies on user-provided keys and browser sessions
- your research needs are purely academic or historical rather than recent-trend focused

## Facets
- artifact type: plugin
- maturity: active
- function: search-engine, agent-framework, rag, web-scraping, nlp
- domain: artificial-intelligence, social-media, developer-tools
- platform: cli, python, cross-platform
- tags: claude-code, agent-skill, deep-research, trend-analysis, reddit, twitter, youtube, hackernews, polymarket, recency, social-listening, ai-agents, search, nodejs

## Member repositories
- mvanhorn/last30days-skill (main) score 82

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:12:18.732452+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T16:18:12.907319+00:00, confidence not recorded.
  - readme: https://github.com/mvanhorn/last30days-skill (fetched 2026-08-28T04:12:18.732452+00:00, sha 52ef01aed2d4)
- Data as of 2026-08-30T08:39:29.467469+00:00.