Weekly LinkedIn drafts for new research papers from your lab or institution

Every Monday at 8am in my timezone, run an agent that finds new research papers published in the previous 7 days from my lab or organization and stages a LinkedIn post draft for each one so my comms team can review before publishing.

Configurable inputs I want to set up front: (1) one or more affiliation strings for my lab/department/institution as they typically appear on papers, e.g. "Smith Lab", "Department of Molecular Biology, Example University"; (2) an optional list of principal investigator names to further narrow results; (3) the LinkedIn author (personal profile or company page) the drafts should be created under.

Step 1 — find new papers. Use Crossref's Search Works operation with query.affiliation set to each configured affiliation string and a from-pub-date filter covering the previous 7 days (today minus 7 through today). If PI names are configured, also pass query.author. Restrict the type filter to research outputs (journal-article and proceedings-article) so we exclude editorials, corrections, retractions, and non-research item types. Use the polite pool (include a mailto). Sort by published descending and page through results until the from-pub-date window is exhausted. De-duplicate by DOI.

Step 2 — filter. Drop any work whose subtype or title indicates a correction, retraction, erratum, withdrawal, or expression of concern. Drop anything without an abstract (we need substance to write from). Keep a small persistent list of DOIs already drafted in earlier runs so we never re-draft the same paper twice.

Step 3 — write a LinkedIn post for each surviving paper. Pull the title, full author list, journal/container-title, DOI, and abstract. Then draft a plain-language post (2 to 4 short paragraphs) that: leads with the practical takeaway in one sentence with no jargon, explains the finding to a non-specialist audience, names the lead author by full name (and tags them if a LinkedIn handle is configured for that PI), names the journal, and ends with the DOI link (https://doi.org/<DOI>). Keep under LinkedIn's 3000-character commentary limit; aim for 1200–1800 characters for readability. Match the tone of a university or biotech comms team: warm, accessible, no academic hedging.

Step 4 — stage each one as a draft on LinkedIn. Use LinkedIn's Create Post operation with the configured author URN and the lifecycleState set to DRAFT (not PUBLISHED), so the post appears in the page's drafts queue for a human to review and publish manually. Use the LinkedIn little-text format and escape reserved characters in the commentary.

Step 5 — be quiet when there's nothing. If the weekly Crossref search returns zero qualifying new works, the run should finish silently with no draft created and no notification. Do not stage placeholder posts.

At the end of each run, record (internally) the DOIs that were drafted this week so the next run's dedupe check works. The output of the workflow is: N LinkedIn drafts staged on the configured page, one per new paper.

Additional information

What does this prompt do?

Looks up papers published in the last 7 days that credit your lab or institution by name.
Pulls the title, authors, journal, DOI, and abstract for each new paper.
Writes a short, jargon-free LinkedIn post that explains the finding to a general audience and tags the lead author.
Stages every post as a draft on your LinkedIn page so a human reviews the copy before anything goes live.
Skips corrections, retractions, and non-research items, and stays quiet on weeks with no new papers.

What do I need to use this?

A LinkedIn account or company page you can post from.
The name of your lab, department, or institution as it appears in paper affiliations.
Optional: a short list of principal investigators or faculty to narrow the results.

How can I customize it?

Change the day and time the search runs, or widen the lookback window beyond one week.
Add the names of specific researchers you want to include, or extra affiliation spellings your team uses.
Tune the writing voice (tone, length, hashtags, calls to action) to match your brand or page style.

Frequently asked questions

Will my LinkedIn posts go live automatically?

No. Every post is staged as a draft on your LinkedIn page so your communications team can read it, edit it, and decide whether to publish.

What if my lab name is written several different ways on papers?

You can list multiple affiliation spellings (full name, common abbreviation, parent institution) and the agent will search for each variant when it pulls new works.

How does it avoid posting about the same paper twice?

The search is scoped to papers published in the previous seven days, and the agent checks against the drafts it created in earlier weeks before staging anything new.

What happens on weeks when no new papers come out?

Nothing. If the weekly search returns no qualifying research articles, the agent stays quiet that week instead of creating an empty draft or a placeholder post.

Can I post from a company page instead of my personal profile?

Yes. When you connect LinkedIn you can choose whether the drafts are staged on a personal profile or a company page you administer.