The release notes baseline selection currently uses octokit.paginate() to fetch all releases, then filters/sorts client-side to find the highest semver tag less than the current version.

For repos with many releases (e.g. 200+), this means multiple API calls to fetch all pages even though the answer is likely on the first page.

Potential optimization

Use octokit.paginate.iterator() to process pages one at a time and short-circuit early once we've confidently found the best candidate:

let previousTag;
for await (const { data: page } of octokit.paginate.iterator(octokit.rest.repos.listReleases, { ... })) {
  // process page, track best candidate
  // break early if we can determine no better candidate exists on later pages
}

Priority

Low — for most repos this is 1-2 API calls total and the current approach is simple and correct. Only worth doing if repos with hundreds of releases become common consumers.