DSS agent skills
Reusable tasks for research with AI agents
A skill gives an agent instructions for a recurring task, together with any scripts or templates it needs. Choose one for the work at hand: setting up a project, finding literature, reviewing an analysis, or running a reproducible workflow.
Install in Claude Code Download all nine skills
Each skill page explains its purpose, requirements, and an example prompt. You direct the research; the agent carries out the task using the access you provide.
Set up and plan
scaffold
Set up a project
Ask the agent to create or adapt folders, project guidance, and runnable commands for a new piece of work.
review-plan
Review a research plan
Find consequential gaps, unsupported assumptions, and missing checks before committing to the work.
dag
Draw a causal diagram
Turn a supplied causal graph into a clear figure, with adjustment sets checked against the graph.
Explore and assess
litsearch
Find research literature
Keep the queries, screening decisions, reading limits, and citation checks behind a literature search.
eda
Explore a dataset
Generate a first-look report on distributions, missing values, and relationships worth investigating.
review-analysis
Review an analysis
Check whether analysis code follows the scientific requirements and supports the quantities it reports.
Run and refine
targets
Build a repeatable R pipeline
Organize an R analysis into stages so completed, unchanged work can be reused when the pipeline runs again.
pueue
Queue computational jobs
Let independent computational jobs run with controlled concurrency while the agent tracks progress and outcomes.
review-prose
Review prose for clarity
Find passages readers have to decode and assess proposed revisions in the surrounding paragraph.
Learn the workflow
Doing Research with AI Agents explains how to organize, check, and document research with an agent. Extending Your Agent’s Toolkit covers the tools behind those workflows.