Add resilient GitHub Pages deploy workflow - #1

The failing deploy job was not caused by site content; actions/deploy-pages@v5 was failing during Pages deployment creation with a transient GitHub 503. This change moves deployment control into a repository-owned workflow and adds bounded retries around the Pages deploy step.

• Replace implicit Pages deployment with an explicit workflow

  • Add .github/workflows/deploy.yml
  • Run on push to main and workflow_dispatch
  • Use the standard Pages flow: checkout → configure Pages → upload artifact → deploy

• Harden deployment against transient Pages API failures

  • Keep Pages deployments serialized with concurrency: pages
  • Disable cancel-in-progress to avoid interrupting active deployments
  • Retry actions/deploy-pages@v5 up to 3 times with short backoff when the first attempt fails

• Stage only publishable site content

  • Copy repository root contents into _site
  • Exclude .git, .github, and _site from the published artifact

• Use current Pages action versions and required permissions

  • actions/checkout@v7
  • actions/configure-pages@v6
  • actions/upload-pages-artifact@v5
  • actions/deploy-pages@v5

concurrency:
  group: pages
  cancel-in-progress: false

- name: Deploy to GitHub Pages
  id: deployment1
  uses: actions/deploy-pages@v5
  continue-on-error: true

- name: Retry GitHub Pages deployment
  if: steps.deployment1.outcome == 'failure'
  uses: actions/deploy-pages@v5