Investigate feasibility of agents.md Dataverse workflow templates for researchers

[cmbz]

Overview

• Researchers who need to share their data in Dataverse repositories are increasingly using AI agents to automate aspects of their data collection-analysis-interpretation-publication workflows.
• Meanwhile, Dataverse installation curators must review deposits to ensure they meet repository requirements.
• Providing agents.md templates for researchers to use in their AI workflows may be labor-saving for researchers and curators, while also ensuring that deposits meet repository standards.
• This issue is for an investigation into the feasibility of developing Dataverse agents.md research workflow templates.

Definition of Done

• Completion of a feasibility study
• Development of several agents.md research workflow templates. These agents.md files will contain instructions for using specific Dataverse APIs and other libraries/modules to construct executable scripts for scanning researcher data files, packaging them, creating Dataverse datasets within a user collection, and uploading the files to the datasets.
• MVP showing these templates in use

Example agents.md:

# Quantitative AI-Ready Data Desposit
## Requirements

- Each data file is comma-delimited or tab-delimited
- Each data file is smaller than 500 GB
- Dataset citation metadata is provided
- Social sciencesmetadata is provided
- Geospatial metadata is provided
- Researcher has data collection
- API key is provided

 ## Actions

- Create sub-collection for dataset
- Create dataset using suppliedmetatdata
- Upload dataset files to sub-collection

## Validation

- Validate deposit using standard method
- Validate deposit using FAIRSCAPE AI-readiness method

## Scoring

- Assign Dataverse completeness score
- Assign FAIRSCAPE AI-readiness score