Natural-language front door: the user gives a journal name or an RSS feed URL; you register it and run the pipeline. Run all commands as PYTHONNOUSERSITE=1 ./env/bin/bio-trend ... from the repo root.

Procedure

Find the feed URL (if only a journal name/homepage was given). Known patterns:
- Nature Portfolio: https://www.nature.com/<code>.rss (journal) or https://www.nature.com/subjects/<slug>.rss (subject).
- Cell Press: https://www.cell.com/<slug>/current.rss (and /inpress.rss).
- Science (AAAS): https://www.science.org/action/showFeed?type=etoc&feed=rss&jc=<code>. Use WebFetch/WebSearch to confirm the exact code/slug if unsure.
Register it — add a row to Journal-RSS.md (Family, Journal, Key, Feed type, RSS URL, Bio focus, Status ?) and a matching entry to config/journals.json (key, label, family, feeds: [{url, type, bio_focus}]). A journal may have multiple feeds.
Verify — bio-trend check-feeds --check. If the new feed is live, update its Status to ✓; if it 403s/dead, mark ✗ and stop (report the problem).
Ingest + refresh — bio-trend ingest --journal <key> then bio-trend refresh.
Smoke-check + PR — confirm data/<key>/ has a CSV and the new journal appears in docs/data/manifest.json. Open a PR summarising the added feed.

Notes

bio_focus: high for biology-dedicated journals, mixed for multidisciplinary flagships (where you may prefer a biology subject feed).
RSS is a rolling window — the first ingest only captures currently-live items; history accrues over subsequent monthly runs.