dataset

What problem does it solve?

This Skill solves the problem of quickly locating and downloading suitable public datasets when a math modeling task requires you to search for external data or benchmark against historical datasets.

Core Features & Use Cases

• Public dataset discovery across Kaggle, UCI, HuggingFace, and Tianchi when the problem explicitly asks for “find similar data” or “supplement external data”.
• Safety guardrail to avoid using this tool when the user’s attachments already contain the needed data (prevents silently fabricating “official” results).
• Standardized landing workflow that saves downloaded datasets into a consistent directory structure and records source/license metadata for auditability.
• Failure diagnostics including Kaggle token guidance, license constraints reminders, large-file handling guidance, and network-restricted fallback instructions.

Quick Start

Tell the agent to find a public dataset matching the topic, download it, and save it under workdir/<task_id>/attachments/external with an accompanying SOURCES.md.