Harden GitHub Pages deploy job with explicit workflow and single retry

[Copilot]

The deploy job was failing after artifact creation with Deployment failed, try again later. in actions/deploy-pages@v5, indicating transient Pages deployment failures rather than a build issue.
This PR introduces a repository-owned Pages workflow that preserves current build behavior and adds retry logic only around deployment.

• Workflow ownership and stability

  • Add .github/workflows/pages.yml to run Pages build/deploy from repository config (push to master + workflow_dispatch).
  • Keep standard Pages permissions and concurrency controls to avoid overlapping deployments.

• Build path parity

  • Build via actions/jekyll-build-pages@v1 and upload via actions/upload-pages-artifact@v4, matching the existing Pages artifact flow.

• Deploy resiliency

  • Run actions/deploy-pages@v5 with a single conditional retry when the first deployment attempt fails.
  • Mark the job failed only if both attempts fail.
  • Surface environment URL from whichever attempt succeeds.

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

- name: Retry deployment once
  id: deploy_retry
  if: steps.deploy_primary.outcome == 'failure'
  uses: actions/deploy-pages@v5