Track specific researchers and labs for new papers, posted to Slack weekly

Every Monday at 8am, run an agent workflow that monitors my Notion watchlist of named researchers and labs for new papers and posts a single ranked Slack digest.

Step 1. Pull the watchlist from Notion. I keep a database called 'Watched researchers' where each row has the researcher's name, their affiliation, an ORCID when I know it, and a short note explaining why I care about them. Use Notion's Query a Data Source operation (not the deprecated Query a Database) to load every active row.

Step 2. For each researcher, call Crossref Search Works with query.author set to their name. When I've recorded an ORCID, add filter=orcid:<orcid> for precision. Always add filter=from-pub-date:<one week ago today> to scope results to the last 7 days. Include my mailto on every Crossref request so we stay in the polite pool.

Step 3. Dedupe against my 'Papers seen' Notion database before doing any further work. That database stores the DOI on every row, so skip anything whose DOI is already logged. This is how next week's run knows what to skip and how we avoid re-posting.

Step 4. For each genuinely new paper, call Crossref Get Work by DOI to fetch the full metadata: title, full author list, container-title (journal), publish date, abstract, and URL.

Step 5. For each new paper, draft a 2 to 3 sentence 'why this matters' blurb that ties the abstract back to the watchlist note I wrote for that author. Be specific. Name the actual connection, not a generic summary of the paper.

Step 6. Post a single Slack message using Send a Message to the channel I configure. Group the digest into two sections, 'Directly competitive' and 'Tangential', based on how closely each paper aligns with my note about that researcher. Within each section, rank by recency. For every paper include the title, authors, journal, publish date, a DOI link, the matched watchlist researcher, and the relevance blurb. Use Slack mrkdwn formatting.

Step 7. Append every new paper to my 'Papers seen' Notion database using Create a Page, with properties for title, authors, journal, DOI, publish date, the relevance blurb, and which watched researcher it was matched to. This is the log next week's run reads in Step 3.

Edge cases. If a watched researcher has no new papers, skip them silently. If none of the watched researchers had new work this week, post a one-line 'no new papers from your watchlist this week' Slack message instead of an empty digest, and don't write anything to the Papers seen log.

Additional information

What does this prompt do?

Reads your watchlist of authors and labs from a Notion database, including the short note you wrote about why each one matters.
Finds every new paper from those researchers published in the last week, using their name and ORCID when you have it.
Skips anything you've already logged, so you only ever see fresh work.
Posts a ranked Slack digest grouped into Directly competitive and Tangential, with a short relevance blurb tying each paper back to your note.
Appends every new paper to a Notion log with title, authors, journal, DOI, and publish date so the record is auditable.

What do I need to use this?

A Notion workspace with two databases shared with General Input: a Watched researchers list and a Papers seen log.
A Slack workspace and the channel you want the digest posted to.
An email address to send with each lookup, so the scholarly metadata service prioritises your requests.

How can I customize it?

Change the schedule. Daily, biweekly, or Friday afternoons work just as well as Monday at 8am.
Send the digest to a private channel or a DM instead of a public channel.
Adjust how the agent groups results. Group by lab, by topic area, or by seniority instead of competitive versus tangential.

Frequently asked questions

Do I need an ORCID for every researcher on my watchlist?

No. The agent uses the ORCID when you provide one for higher-precision matching, and falls back to a name and affiliation search when you don't.

What if two of my watched researchers co-author the same paper?

The agent dedupes by DOI so you never see the same paper twice, but it credits the paper to every watched author who appears on it and calls out the collaboration in the blurb.

Will this catch preprints from arXiv or bioRxiv?

The default source is Crossref, which covers what publishers deposit. Most preprint servers are not indexed there. If you also want preprints, ask the agent to add a preprint source as a secondary check.

How does it decide what's Directly competitive versus Tangential?

It uses the short note you wrote next to each researcher in the Notion watchlist. The agent compares the paper's abstract to that note and ranks based on how closely they overlap.

Can I share the Notion log with my team?

Yes. The Papers seen database is a regular Notion page, so you can share it with teammates or embed it in another doc like any other Notion content.