Fix failing update-citations workflow (timeout, NameError, git add crash)

[Copilot]

The update-citations CI job was failing with two cascading errors: the scholarly fetch timed out after 90s (exit 124), then git add _data/citations.yml crashed with exit 128 because the file was never created.

Root causes fixed:

• NameError in bin/update_scholar_citations.py: existing_data was assigned only inside if os.path.exists(OUTPUT_FILE) but referenced unconditionally at line 112. Initialized to None before the block.
• git add exit 128: git add _data/citations.yml fails when the file doesn't exist (script timed out or errored). Added a file-existence guard — skips commit step gracefully if no file present.
• Timeout too short: Bumped timeout from 90s → 300s; scholarly makes multiple retried network requests to Google Scholar and reliably exceeded 90s.

Ownership Routing

• I confirmed this PR targets the correct repo based on docs/BOUNDARIES.md.
• This PR only changes starter-owned scope (al-folio) or I am porting a routed change and linked the owning repo issue/PR.

Owning repo (if not starter):
Related issue/PR:

Plugin Ecosystem (if applicable)

• Not applicable

Starter Wiring Changes

• Not applicable

Validation

• npm ci
• bundle exec jekyll build
• npm run lint:prettier
• npm run lint:style-contract
• Integration tests (test/integration_*.sh) as needed
• Visual tests (npm run test:visual) as needed

Notes

The underlying scholarly network calls may still be flaky (Google Scholar rate-limits scrapers). If timeouts persist at 300s, consider adding a free proxy via scholarly.use_proxy() or caching results more aggressively.