pubmed-daily-digestlisted

Daily medical-literature digest agent — searches the past 24 hours on PubMed across configurable thematic queries, fills in missing abstracts via CrossRef, writes a TL;DR and a one-line Hot Take for each article (English by default; language is configurable in Step 2/3), picks 3–5 notable highlights, and renders a Markdown digest to a local output directory. Trigger when the user says "run daily digest", "daily lit", "daily journal", "medical literature digest", or "跑每日文獻". For yearly maintenance the skill also handles "update SJR scores" / "refresh journal rankings" — see the Maintenance section near the bottom.
didiowen/pubmed-daily-digest · ★ 0 · AI & Automation · score 70

Install: claude install-skill didiowen/pubmed-daily-digest

# PubMed Daily Digest > A Claude Code skill that pulls fresh PubMed articles every day across a few > thematic queries, refills missing abstracts via CrossRef, summarises each in > English, picks 3–5 notable highlights, and writes a Markdown digest to a > local output directory. Publishing the digest anywhere is up to you. ## Configuration Edit these values in this file before running in a new setup: | Key | Default | Notes | |--------------------|------------------------------------------|-------------------------------------------------------------| | `OUTPUT_DIR` | `output/` (inside this skill folder) | Where `{DATE}.md`, `{DATE}_raw.json`, `{DATE}_daily.json` go | | `TIMEZONE` | `UTC` | IANA name passed to `ZoneInfo` in Step 0 | | `CROSSREF_MAILTO` | `pubmed-daily-digest@noreply.example` | CrossRef polite-pool identifier | | `QUERIES` | Four thematic queries, see Step 1 table | Customise freely — see `README.md` for guidance | Scripts in `scripts/` resolve `sjr_curated.json` and the rolling `.seen_pmids.json` cache relative to their own path, so they work from any cwd as long as the folder layout is intact. **Before the first run, build your own `sjr_curated.json`.** The repo ships `sjr_curated.example.json` only as a schema reference — its c