Skip to content
SkillScript-led6,675 tokens in use

last30days

Research any topic from the last 30 days across Reddit, X, YouTube, TikTok, Hacker News, Polymarket, GitHub and the web, then get a cited summary.

71,289 stars on the repository MIT Licensemvanhorn/last30days-skill · skills/last30days
npx skills add mvanhorn/last30days-skill -g
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing last30days costs

59tokens idle
6,675tokens in use
73,667tokens, up to

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Up to adds the files its instructions link. Counts are approximate, made offline with cl100k_base.

Model dependence Script-led

Script-ledMixedModel-led

Most of this skill is scripts, so it behaves much the same whichever model runs it.

This describes the skill's structure, how much of it is scripts. It does not say how well it works. Scripts: scripts/box_chrome_login.py, scripts/briefing.py, scripts/build-skill.sh, scripts/compare.sh, scripts/evaluate_search_quality.py, scripts/last30days.py, scripts/lib/__init__.py, scripts/lib/agentcookie.py, scripts/lib/amazon.py, scripts/lib/arxiv.py, scripts/lib/backends.py, scripts/lib/bird_x.py, scripts/lib/bluesky.py, scripts/lib/bounded_get.py, scripts/lib/brightdata.py, scripts/lib/categories.py, scripts/lib/chrome_cdp.py, scripts/lib/chrome_cookies.py, scripts/lib/cjk.py, scripts/lib/cluster.py, scripts/lib/cluster_diversity.py, scripts/lib/company_names.py, scripts/lib/company_resolve.py, scripts/lib/competitors.py, scripts/lib/cookie_extract.py, scripts/lib/cookie_paths.py, scripts/lib/corpus.py, scripts/lib/dates.py, scripts/lib/dedupe.py, scripts/lib/deepline.py, scripts/lib/devto.py, scripts/lib/digg.py, scripts/lib/discovery_handoff.py, scripts/lib/doctor.py, scripts/lib/dripstack.py, scripts/lib/employer_site.py, scripts/lib/entity_extract.py, scripts/lib/env.py, scripts/lib/fanout.py, scripts/lib/feed.py, scripts/lib/fetch_memo.py, scripts/lib/freshness.py, scripts/lib/fusion.py, scripts/lib/github.py, scripts/lib/grok_audit.py, scripts/lib/grok_x.py, scripts/lib/grounding.py, scripts/lib/hackernews.py, scripts/lib/health.py, scripts/lib/hiring_signals.py, scripts/lib/hosted.py, scripts/lib/html_publish.py, scripts/lib/html_render.py, scripts/lib/http.py, scripts/lib/instagram.py, scripts/lib/job_boards/__init__.py, scripts/lib/job_boards/detect.py, scripts/lib/job_boards/google.py, scripts/lib/job_boards/ibm.py, scripts/lib/job_boards/jibe.py, scripts/lib/job_boards/kula.py, scripts/lib/job_boards/oracle.py, scripts/lib/job_boards/phenom.py, scripts/lib/job_boards/robots.py, scripts/lib/job_boards/sitemap.py, scripts/lib/job_boards/talentbrew.py, scripts/lib/job_boilerplate.py, scripts/lib/jobs.py, scripts/lib/jobs_digest.py, scripts/lib/jobs_story.py, scripts/lib/jobs_teams.py, scripts/lib/library.py, scripts/lib/library_index.py, scripts/lib/linkedin.py, scripts/lib/log.py, scripts/lib/meta_ads.py, scripts/lib/normalize.py, scripts/lib/parallel_mcp.py, scripts/lib/patent_digest.py, scripts/lib/patents.py, scripts/lib/permission_preflight.py, scripts/lib/perplexity.py, scripts/lib/person.py, scripts/lib/person_linkedin.py, scripts/lib/pinterest.py, scripts/lib/pipeline.py, scripts/lib/planner.py, scripts/lib/polymarket.py, scripts/lib/preflight.py, scripts/lib/prescriptions.py, scripts/lib/private_files.py, scripts/lib/providers.py, scripts/lib/proxy.py, scripts/lib/public_host.py, scripts/lib/quality_nudge.py, scripts/lib/query.py, scripts/lib/reddit.py, scripts/lib/reddit_arctic.py, scripts/lib/reddit_discover.py, scripts/lib/reddit_enrich.py, scripts/lib/reddit_envelope.py, scripts/lib/reddit_health.py, scripts/lib/reddit_hydrate.py, scripts/lib/reddit_keyless.py, scripts/lib/reddit_listing.py, scripts/lib/reddit_public.py, scripts/lib/reddit_refs.py, scripts/lib/reddit_search.py, scripts/lib/reddit_session.py, scripts/lib/reddit_shreddit.py, scripts/lib/registers.py, scripts/lib/relevance.py, scripts/lib/render.py, scripts/lib/report_checklist.py, scripts/lib/rerank.py, scripts/lib/resolve.py, scripts/lib/safari_cookies.py, scripts/lib/schema.py, scripts/lib/setup_wizard.py, scripts/lib/signals.py, scripts/lib/skill_meta.py, scripts/lib/snippet.py, scripts/lib/sqlite_snapshot.py, scripts/lib/stocktwits.py, scripts/lib/subproc.py, scripts/lib/techmeme.py, scripts/lib/telegram.py, scripts/lib/threads.py, scripts/lib/tiktok.py, scripts/lib/topic_shape.py, scripts/lib/transcribe.py, scripts/lib/trustpilot.py, scripts/lib/truthsocial.py, scripts/lib/ui.py, scripts/lib/usage.py, scripts/lib/vendor/bird-search/bird-search.mjs, scripts/lib/vendor/bird-search/lib/cookies.js, scripts/lib/vendor/bird-search/lib/paginate-cursor.js, scripts/lib/vendor/bird-search/lib/runtime-features.js, scripts/lib/vendor/bird-search/lib/runtime-query-ids.js, scripts/lib/vendor/bird-search/lib/twitter-client-base.js, scripts/lib/vendor/bird-search/lib/twitter-client-constants.js, scripts/lib/vendor/bird-search/lib/twitter-client-features.js, scripts/lib/vendor/bird-search/lib/twitter-client-search.js, scripts/lib/vendor/bird-search/lib/twitter-client-types.js, scripts/lib/vendor/bird-search/lib/twitter-client-utils.js, scripts/lib/web_fetch_keyless.py, scripts/lib/web_search_keyless.py, scripts/lib/windows_job.py, scripts/lib/x_api.py, scripts/lib/x_envelope.py, scripts/lib/x_judge.py, scripts/lib/xai_x.py, scripts/lib/xiaohongshu_api.py, scripts/lib/xquik.py, scripts/lib/xurl_x.py, scripts/lib/yahoo_finance.py, scripts/lib/youtube_yt.py, scripts/reddit_canary.py, scripts/setup-keychain.sh, scripts/setup-pass.sh, scripts/store.py, scripts/test-v1-vs-v2.sh, scripts/test_device_auth.py, scripts/verify_v3.py, scripts/watchlist.py, scripts/x_lane_parity.py.

Touches

shelldeclared and detectednetworkdeclared and detectedwrites filesdeclared and detectedcredentialsdetected

Declared tags come from the skill's frontmatter; detected ones from its scripts and code blocks. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Good
  • Frontmatter11/15
  • Structure19/25
  • Examples9/15
  • Output15/15
  • Size13/15
  • Resources1/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Needs attention
  • File that was not read−15

    a file the skill ships that was not read, so what it does is unknown

    .skillignore
  • Reads credentials−15

    does this on the machine, and the frontmatter does not say so

    SKILL.md, line 132
  • Runs shell commands−2

    does this on the machine, as the frontmatter declares

    SKILL.md, line 132
  • Writes files−4

    does this on the machine, as the frontmatter declares

    scripts/box_chrome_login.py, line 194
  • Reaches the network−5

    does this on the machine, as the frontmatter declares

    scripts/evaluate_search_quality.py, line 18
  • Hidden characters−20

    characters that render as nothing or reorder the text

    scripts/lib/render.py, line 3547

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Excellent
  • Development activity100
  • Maintenance100
  • Community68
  • Maturity32
  • Momentum40

last30days is an agent skill that researches what people say about a topic in the last 30 days. It pulls posts and engagement from Reddit, X, YouTube, TikTok, Hacker News, Polymarket, GitHub and the web, scores the results by upvotes, likes and market odds, and has the agent write one brief.

It runs a Python engine from the skill folder. Reddit, Hacker News, Polymarket and GitHub work without configuration, and a setup wizard on first run unlocks more sources. It installs through the Claude Code plugin marketplace or the npx skills CLI.

What You Get

  • A research brief with inline links to the posts, threads and videos behind each claim
  • Evidence from Reddit, X, YouTube, TikTok, Hacker News, Polymarket, GitHub and web search, weighted by engagement
  • A doctor health check that diagnoses broken or missing sources
  • Comparison reports for two or three topics, with a quick verdict and a head-to-head section
  • Discovery of trending topics, an offline library of past runs, and optional HTML export

Common Use Cases

  • Reading the last month of posts and talks about a person before a meeting
  • Checking what a company is hiring for and filing patents on before a sales call
  • Finding the prompts and techniques a community has settled on for a new AI tool
  • Comparing two products by what their users say on Reddit and X
  • Tracking which topics are trending in a domain this month

What Makes It Unique

The skill ranks evidence by what people do: Reddit upvotes, X likes, YouTube transcripts, and Polymarket odds backed by money. It works across many walled platforms in one run, keeps a 30-day window, and has the host agent act as planner and judge instead of calling a separate judging service. It also covers hiring signals, patent filings, comparisons, follow-up questions on saved research and a source health check.

Under The Hood

The SKILL.md is a router. It tells the agent which reference file to read for each kind of request, such as runtime, setup wizard, research runbook, synthesis, comparison or discovery. The agent then runs the Python 3.12+ engine in scripts/ through the shell. The engine calls the services of each enabled source over the network, using optional API keys and tokens that you set, and can read a logged-in Reddit or X browser session only after you consent. The frontmatter lists the API key names it may use, but a security scan found that the skill reaches credentials on the machine (keychain and pass helpers, browser cookies and tokens) without declaring that in its frontmatter allowed tools. The scan also flagged invisible or text-reordering characters in scripts/lib/render.py, and could not read the .skillignore file the skill ships, so what that file does was not checked. The engine stores a config file under ~/.config/last30days, keeps a local SQLite library of runs, and saves reports to ~/Documents/Last30Days unless you change that. First-run setup can install free command-line tools such as arxiv-pp-cli and techmeme-pp-cli. The frontmatter allows Bash, Read, Write, AskUserQuestion and WebSearch.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers