fix(ci): replace fragile two-job Pages deployment with single-job workflow

[Copilot]

The automatic pages build and deployment workflow splits artifact upload (build) and deployment (deploy) into separate jobs. When the deploy job fails for any transient reason and only that job is re-run, the ephemeral github-pages artifact (uploaded by build, ~24h retention) is already gone — every subsequent retry fails with "No artifacts named 'github-pages' were found", requiring a full "Re-run all jobs" to recover.

Changes

• Added .github/workflows/pages-deploy.yml — custom workflow triggered on push to gh-pages that combines checkout, upload-pages-artifact, and deploy-pages in a single job. Every run (including re-runs) always produces a fresh artifact immediately before deploying, eliminating the expiry window entirely.

Required repository setting change

This workflow requires Pages source set to GitHub Actions (Settings → Pages → Source → GitHub Actions). Once set, this workflow replaces the automatic two-job workflow entirely.

jobs:
  deploy:
    environment:
      name: github-pages
      url: ${{ steps.deployment.outputs.page_url }}
    steps:
      - uses: actions/checkout@v6
        with: { ref: gh-pages }
      - uses: actions/upload-pages-artifact@v3
        with: { path: . }
      - uses: actions/deploy-pages@v5        # ← always has a fresh artifact