github-activity-report

Use the gh CLI to list all non-archived org repos via GraphQL with pagination, then iterate each repo's pulls and commits endpoints for the chosen date window. Aggregate the JSONL output with jq to rank authors and repos.

The search API paginates only the first 1,000 results even when total_count is higher, so per-author and per-repo splits are biased toward GitHub's sort order. Crawling each repo's pulls and commits endpoints directly avoids this cap entirely.

Yes. The scope picker lists all orgs you belong to plus your personal account. For personal scope it queries viewer repositories with OWNER affiliation, excluding forks and archived repos.

Crawls typically take 5 to 20 minutes depending on org size, since each repo requires paged calls to two endpoints. The script runs in the background and logs per-repo progress that can be monitored every 30 seconds.

LOC stats require one API call per commit to fetch additions and deletions, which means thousands of requests for an active org and triggers secondary rate limits. They are only attempted if explicitly requested.

Bot accounts such as dependabot[bot] are included in the raw rankings and can dominate them. The report flags this in its caveats section, and the raw JSONL data can be re-filtered to produce a humans-only view.