|
|
|
K-Dense Updates
// May 29, 2026
|
|
Our newsletter took a little vacation last week as we celebrated the long weekend, but we’re back at it this week with some exciting updates.
|
>Thank You, Community!
Our Scientific Agent Skills repo crossed 25,000 stars on GitHub last week. 🎉
Thank you to every researcher, builder, and contributor. Your issues, PRs, and feedback are what make Scientific Agent Skills the most comprehensive open library of scientific skills available today. We’re incredibly grateful to be building this with you.
If Scientific Agent Skills is useful in your work, please star the repo on GitHub — it helps more scientists discover the project and keeps us building.
|
>Agentic Scientist Summit Workshop
We’re hosting a 2-hour workshop at The Agentic Scientist Summit, taking place June 16–17, 2026 in San Francisco, CA. Workshop details forthcoming.
View summit info →
|
>New Scientific Agent Skills
We have four new skills added to the Scientific Agent Skills repo:
|
•LiteParse (from the LlamaIndex team) allows your AI co-scientist to parse PDFs, extract data, run OCR on scanned documents, and more.
|
|
•Nextflow + nf-core enables your agent to run community pipelines, write and nf-test modules/subworkflows, configure Docker/Singularity/Conda, scale to SLURM or the cloud, and debug using best practices.
|
|
•pathway-enrichment equips your agent to run gene-set and pathway enrichment analysis, surfacing the biological processes behind a list of genes.
|
|
•bulk-rnaseq guides your agent through bulk RNA-seq analysis, from quality control through differential expression.
|
We also updated over 60 of our current skills this week.
|
>New Release: Science Superpowers
We just released Science Superpowers, a complete computational-science methodology for your research agents, built on a set of composable skills plus initial instructions that make sure your agent actually uses them.
It’s a science-domain reimplementation of Superpowers, swapping test-driven development for pre-registration: the skills auto-trigger at session start to turn a fuzzy interest into a falsifiable question, ground it in prior work, lock hypotheses and decision rules before looking at outcomes, then execute, verify, and red-team the result in a reproducible workspace.
|
>What’s New in K-Dense Web
A little UI refresh from the following week brought the following updates:
|
•Session timers removed
|
|
•Question and planning responses are expanded by default
|
|
•Line breaks respected in user input
|
|
•Added real-time user feedback panel to right side of window
|
|
>Stay Connected
Follow us for updates between newsletters on
X,
LinkedIn, and
YouTube.
|
|
Until next week,
The K-Dense Team
|
|
|
K-Dense Inc.
380 Portage Avenue
Palo Alto, CA 94306
|
Unsubscribe
|
|
|