inno-prepare-resources

Automating research resource preparation involves loading an evaluation instance, searching GitHub for related repositories, and downloading arXiv paper sources. This Skill orchestrates metadata, references, and dataset prompts to curate reference codebases and papers.

Gathering reference codebases and datasets requires coordinating resources for both Idea mode and Plan mode. This Skill derives a date_limit from an evaluation instance to guide dataset selection and aggregates GitHub repository search results for curation.

Downloading arXiv paper sources builds a reference corpus by retrieving paper data linked to your evaluation instance. The Skill coordinates with the arXiv service to fetch sources needed for publication-ready artifacts and experiments.

Automating your research workflow requires a properly formatted SKILL.md frontmatter. This file must include name and description fields, along with path values for Ideation.references, Experiment.code_references, and Experiment.datasets.

Searching GitHub for repositories related to evaluation papers is a core feature. The Skill loads your instance and automatically searches GitHub to discover and aggregate relevant repositories for your research curation.

Orchestrating metadata and dataset prompts requires access to the Prepare Agent and arXiv services. Without the correctly formatted SKILL.md frontmatter specifying paths for references, code, and datasets, the automation cannot curate the resources.