GitHub Repo Intelligence: Vet Any Dependency Before You Add It

The GitHub UI shows you a star count and a green “Active” label on the repo page, and neither one tells you if a project is actually maintained. What matters is when the last commit landed, whether releases are still shipping, and how many people beyond the original author are contributing. That’s several tabs to check by hand for every candidate dependency, which is exactly why most people check none of them.

Try it live: GitHub Repo Intelligence. Pay only for results delivered, no result no charge.

The GitHub Repo Intelligence actor reads GitHub’s public API and returns the whole picture for a repo in one call, no browser, no token.

What you get back

Pass a list of repos (owner/name strings) and an includeReadme flag. Each repo comes back as one record:

Field	Notes
stars, forks, openIssues	Current counts
primaryLanguage, languages	Language breakdown by bytes
contributorsCount	People beyond the original author
latestReleaseTag, latestReleaseAt	Most recent tagged release
lastCommitAt, createdAt, updatedAt	The three dates that actually tell you if it’s alive
isArchived, isFork	Two flags worth checking before anything else
readme	Full text, if includeReadme is true

A real pull

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")

run = client.actor("themineworks/github-repo-intelligence").call(run_input={
    "repos": ["anthropics/claude-code"],
    "includeReadme": True,
})

for row in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(row["fullName"], row["stars"], row["contributorsCount"], row["latestReleaseTag"])

Real output for anthropics/claude-code: 141,758 stars, 22,761 forks, 55 contributors, latest release v2.1.233 shipped August 14, 2026, repo created February 2025. Not archived, not a fork. That’s the profile of an actively maintained project, and it’s the same four fields you’d check for a much smaller, much riskier dependency.

How to read it

Last commit date first. If it’s months old on a library you’re about to depend on, that’s most of the answer before you look at anything else.

Contributors, not stars. Stars measure who noticed a project once. Contributor count measures whether more than one person can actually keep it alive if the original author moves on.

Release cadence against open issue count. A repo with hundreds of open issues and no release in a year is a repo where your bug report joins a queue nobody is clearing.

Honest limits

Commit activity and release cadence are signals, not proof. Some genuinely stable libraries go quiet because the job is done, not because they’ve been abandoned. Read the description and issue count alongside the dates before ruling anything out on commit recency alone.

Pricing

$0.002 per repository analyzed on the free plan, down to $0.0012 on higher tiers. A repo that fails to resolve is not charged.

Use it from an AI agent

https://mcp.apify.com/?tools=themineworks/github-repo-intelligence

Related

• GitHub Skill & Memory Discovery for finding agent-config repos across GitHub, not just checking one you already have
• RAG Crawler for turning a repo’s docs site into clean markdown once you’ve decided to depend on it
• SEC EDGAR Full-Text Search for the same “public data, one clean call, no auth” shape applied to company filings