带脚本的论文摘要生成

Overview

Use this skill when arXiv paper URLs or local paper files need deterministic preprocessing before the four-stage paper summarization workflow runs. The bundled script downloads arXiv PDFs to local storage when URLs are provided, then handles extraction and cleaning. Do not parse paper web pages or use HTML content as the paper source. After preprocessing, run three independent generation stages from the same cleaned paper text, then a fourth verification stage that evaluates all three generated outputs against the original text.

Canonical inputs

Normalize the request into:

• language
• paperurls for arXiv inputs
• paperfiles

Treat empty string, [], null, None, missing field, or blank list as empty.