delete-stale-branches.yml

# Runs daily to clean up two categories of branches:
#
#   1. Branches whose PR has been closed (merged or otherwise) — acts as a
#      retroactive safety net for anything the event-driven workflow missed.
#
#   2. Branches with no associated PR at all that haven't had a commit in
#      over 30 days — these are abandoned branches that were never turned
#      into a PR.
#
# Both parts respect protected branches and skip fork branches.
# Up to 100 branches are deleted per run, oldest first.
#
# To test without deleting anything, trigger this workflow manually via
# workflow_dispatch and leave dry-run set to "true" (the default). The
# run will log "Would delete: ..." for every branch it would have removed.

name: Delete stale branches

on:
  schedule:
    - cron: "0 4 * * *"
  workflow_dispatch:
    inputs:
      dry-run:
        description: "Dry run (log what would be deleted without deleting)"
        required: false
        default: "true"
        type: choice
        options:
          - "true"
          - "false"

permissions:
  contents: write
  pull-requests: read

jobs:
  delete-stale-branches:
    runs-on: ubuntu-latest
    steps:
      - name: Delete closed PR branches and unassociated branches older than 90 days
        env:
          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
          DRY_RUN: ${{ github.event_name == 'workflow_dispatch' && inputs.dry-run || 'false' }}
        run: |
          LIMIT=100
          deleted=0

          if [ "$DRY_RUN" = "true" ]; then
            echo "*** DRY RUN — no branches will be deleted ***"
          fi

          protected=$(gh api "repos/$GITHUB_REPOSITORY/branches?protected=true&per_page=100" \
            | jq -r '.[].name')

          is_protected() {
            echo "$protected" | grep -qx "$1"
          }

          delete_branch() {
            local branch="$1"
            local reason="$2"

            if is_protected "$branch"; then
              echo "Skipped (protected): $branch"
              return
            fi

            if [ "$DRY_RUN" = "true" ]; then
              echo "Would delete ($reason): $branch"
              deleted=$((deleted + 1))
              return
            fi

            gh api "repos/$GITHUB_REPOSITORY/git/refs/heads/$branch" \
              -X DELETE 2>/dev/null \
              && { echo "Deleted ($reason): $branch"; deleted=$((deleted + 1)); } \
              || echo "Skipped (already gone): $branch"
          }

          echo "--- Closed PR branches (oldest first, limit $LIMIT) ---"

          gh pr list \
            --repo "$GITHUB_REPOSITORY" \
            --state closed \
            --limit 500 \
            --json headRefName,headRepositoryOwner,baseRepositoryOwner,closedAt \
            | jq -r '
                .[]
                | select(.headRepositoryOwner.login == .baseRepositoryOwner.login)
                | [.closedAt, .headRefName]
                | @tsv
              ' \
            | sort \
            | awk -F'\t' '{print $2}' \
            | sort -u \
            | while read -r branch; do
                [ "$deleted" -ge "$LIMIT" ] && break
                delete_branch "$branch" "closed PR"
              done

          echo "--- Unassociated branches older than 30 days (oldest first, limit $LIMIT) ---"

          pr_branches=$(gh pr list \
            --repo "$GITHUB_REPOSITORY" \
            --state all \
            --limit 1000 \
            --json headRefName,headRepositoryOwner,baseRepositoryOwner \
            | jq -r '.[] | select(.headRepositoryOwner.login == .baseRepositoryOwner.login) | .headRefName' \
            | sort -u)

          cutoff=$(date -d '30 days ago' --utc +%Y-%m-%dT%H:%M:%SZ 2>/dev/null \
            || date -u -v-30d +%Y-%m-%dT%H:%M:%SZ)

          echo "Cutoff date: $cutoff"

          candidates=""

          page=1
          while true; do
            batch=$(gh api "repos/$GITHUB_REPOSITORY/branches?per_page=100&page=$page" \
              | jq -r '.[].name')
            [ -z "$batch" ] && break

            while read -r branch; do
              if echo "$pr_branches" | grep -qx "$branch"; then
                continue
              fi

              last_commit=$(gh api "repos/$GITHUB_REPOSITORY/commits?sha=$branch&per_page=1" \
                | jq -r '.[0].commit.committer.date')

              if [[ "$last_commit" < "$cutoff" ]]; then
                candidates="$candidates$last_commit\t$branch\n"
              fi
            done <<< "$batch"

            page=$((page + 1))
          done

          printf "$candidates" | sort | awk -F'\t' '{print $2}' | while read -r branch; do
            [ "$deleted" -ge "$LIMIT" ] && break
            delete_branch "$branch" "no PR, last commit $(printf "$candidates" | grep -F "$branch" | awk -F'\t' '{print $1}')"
          done

          echo "--- Done: $deleted branch(es) ${DRY_RUN:+would be }deleted ---"