setup-repo

Arguments

Parse $ARGUMENTS for:

• Positional argument: default branch name override (e.g. main, master)
• --checks <name1,name2,...>: explicitly specify required status checks (skip auto-detection)

Examples:

• /setup-repo → auto-detect everything
• /setup-repo main → protect the main branch
• /setup-repo --checks "Lint,Test,Notify PR" → explicit checks, skip auto-detection

Step 1: Identify repo and default branch

Run:

gh repo view --json nameWithOwner,defaultBranchRef --jq '{repo: .nameWithOwner, branch: .defaultBranchRef.name}'

• Extract REPO (e.g. lexfrei/kuberture) and BRANCH (e.g. master)
• If a positional argument was provided, use it as BRANCH instead

Report to the user: "Configuring {REPO}, protecting branch {BRANCH}"

Step 2: Configure repo merge settings

Run:

gh api --method PATCH repos/{REPO} \
  --field allow_squash_merge=true \
  --field allow_merge_commit=false \
  --field allow_rebase_merge=false \
  --field delete_branch_on_merge=true

Report: "Repo settings: squash-only merges, auto-delete branches"

Step 3: Detect required status checks

If --checks was provided: parse the comma-separated list and skip to Step 4.

Otherwise, auto-detect from workflow files:

1. Find all workflow files: .github/workflows/*.yml and .github/workflows/*.yaml

2. For each workflow file, read it and check if it has a pull_request trigger in the on: section. Skip workflows that only trigger on push, release, schedule, etc.

3. For each PR-triggered workflow, analyze the jobs: section:

   • Build a map of job_key → { name: <display name>, needs: [<dependencies>] }
   • If a job has no name: field, use the job key as the display name
   • Collect all job keys that appear in ANY other job's needs: list — these are "depended-upon" jobs
   • "Terminal jobs" are jobs whose key does NOT appear in any other job's needs: list
   • The required check names are the name: fields of the terminal jobs