Weekly research paper digest for Slack (Crossref + Google Sheets)

Build me a weekly literature radar that runs on a cron schedule every Monday at 8:00am in my local timezone.

On each run, the agent should do the following:

1. Compute the date range for the previous seven days and use it as from-pub-date and until-pub-date in every Crossref call.

2. Call Crossref Search Works with query.bibliographic set to my topic keywords [TOPIC KEYWORDS HERE], filtered by the seven-day window, sorted by published descending, with rows=100. Use the polite pool by passing a mailto query param with my email.

3. Call Crossref List Works in Journal once per ISSN on my watchlist [JOURNAL ISSNS HERE], filtered by the same seven-day window, sorted by published descending, with rows=100.

4. Merge both result sets and dedupe by DOI.

5. Read the prior rows from my Google Sheet [SPREADSHEET ID + TAB NAME] and drop any merged item whose DOI already appears in the sheet. This is the cross-week dedupe so the same preprint never gets posted twice.

6. Rank the surviving items by relevance to my topic keywords and likely importance. Consider journal reputation, author signals, novelty of the contribution, and how well the title and abstract match my interests. Pick the top eight.

7. For each pick, draft a one or two sentence annotation explaining the contribution and why it matters to someone working in this field.

8. Post a single digest message to Slack channel [CHANNEL ID OR NAME] using the Slack Send a Message operation. Format the message in Slack mrkdwn with a header line for the week, then one block per paper containing the bolded title, the author list, the journal name, the publication date, the agent's annotation, and a clickable DOI link of the form https://doi.org/<DOI>.

9. Append one row per chosen paper to my Google Sheet using the Google Sheets Append Values operation with these columns in order: run_date, doi, title, authors, journal, publication_date, annotation.

If, after dedupe, no papers remain to post, send a short Slack message to the same channel saying nothing notable came through this week so the team knows the run completed.

Keep all topic keywords, journal ISSNs, the Slack channel, the Google Sheet ID, and the polite-pool email as configurable inputs at the top of the workflow so I can edit them without touching the rest of the logic.

Additional information

What does this prompt do?

Searches the world's scholarly literature for new papers published in the last seven days that match your topics and watchlist of journals.
Ranks the matches by relevance and likely importance, then writes a short one or two sentence annotation explaining each pick.
Posts the top eight to a Slack channel of your choice with title, authors, journal, publication date, and a clickable DOI link.
Appends every pick to a Google Sheet so you build a longitudinal reading log and the agent never surfaces the same paper twice.

What do I need to use this?

A Slack workspace and a channel for the weekly digest.
A Google account with a Sheet where the agent can append rows.
A short list of topic keywords and the journals (by ISSN) you want to track.

How can I customize it?

Change the schedule. Run it daily, every Friday, or twice a week instead of Monday at 8am.
Adjust the topic keywords, journal watchlist, or how many papers land in each digest.
Tune what the agent favors when ranking. For example, prioritize open access, specific publishers, or first-time authors.

Frequently asked questions

How does it pick which papers to surface?

The agent ranks last week's matches against your topic keywords and likely importance signals like journal reputation, author affiliations, and novelty, then picks the top eight.

Will the same preprint show up twice?

No. Before posting, the agent checks the Google Sheet log from prior weeks and drops anything already covered, even if it shows up again in a different journal.

Do I need a Crossref account or paid plan?

No. Crossref's scholarly metadata is a public service. There is no signup, no API key, and no cost.

Can I track more than one research area?

Yes. Add as many topic keyword groups and journal ISSNs as you want. The digest will draw from all of them and dedupe before posting.

What happens if no new papers match in a given week?

The agent posts a short note to your Slack channel saying nothing notable came through that week, so you always know it ran.