Add one visible retry to remote E2E tests

[isaacroldan]

Why

Remote E2E failures currently have no controlled retry. This makes transient failures look the same as persistent product failures.

What

• Enable one retry for the remote project in CI only.
• Keep local tests and remote authentication without retries.
• Report first-attempt failures, passed retries, and persistent failures separately.
• Include the retry index when generating app and store timestamps so each attempt has distinct cleanup targets.
• Give the remote Playwright run 30 minutes inside a 40-minute CI job.

Testing

• pnpm --filter e2e lint
• pnpm --filter e2e type-check
• pnpm --filter e2e exec playwright test tests/retry-behavior.spec.ts --project local
• pnpm --filter e2e exec playwright test --project local
• Loaded the retry reporter with CI configuration.
• Verified local and remote-auth use zero retries and remote uses one retry.
• Parsed tests-pr.yml as YAML.