fix: make content refresh checkpoint self-healing

[hanihusam]

Summary

• refresh every tracked content file when the deployed refresh checkpoint is missing
• advance the checkpoint after every successful deploy, including deploys with no content changes
• fail the workflow on git, HTTP, or timeout errors instead of reporting a false success
• target the actual staging Fly app when refreshing dev

Why

After PR #162 deployed, the refresh job read the new build SHA as its comparison SHA. It therefore compared the deployment with itself, found no changed content, and left the projects directory cache stale. The new case study returned 404 even though its commit was running in production.

The first deployment of this fix will bootstrap the missing checkpoint by refreshing all tracked MDX files. Later deployments will use incremental diffs from the stored checkpoint.

Verification

• focused Node harness covering missing-checkpoint bootstrap, incremental content changes, and unchanged deployments
• staging hostname assertion
• npm run validate

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud