← ClaudeAtlas

summary-paperslisted

Process PDF papers supplied as files, folders, or path lists; recursively discover folder contents, extract text and optional images, summarize every paper into structured JSON, then export Markdown, Excel, HTML, or all outputs. Use for batch PDF-paper summaries, including papers stored in different folders.
seven0813/Summary-Paper-Skill · ★ 10 · Data & Documents · score 74
Install: claude install-skill seven0813/Summary-Paper-Skill
# Summary-Papers Use this skill when the user wants to batch-read local PDF papers and export paper summaries as Markdown, Excel, HTML, or all formats. Inputs may mix individual PDF paths, directories, and UTF-8 path-list files; never copy PDFs into a staging folder just to process them. ## Required user input Ask for or confirm these values before doing any work: ```text paper_inputs: - /absolute/path/to/paper.pdf - /absolute/path/to/paper/folder input_lists: optional UTF-8 text files containing one paper input per line output_root: /absolute/path/to/output/folder output: markdown | excel | html | all ``` Interpret the input strictly: - Each paper input must be an existing PDF file or local directory. - Recursively scan directory inputs and ignore their non-PDF contents. - A path-list file contains one PDF or directory per line. Ignore blank lines and lines beginning with `#`; resolve relative entries from the list file's directory. - Resolve and deduplicate PDFs before processing. Read source PDFs in place; do not copy, move, or rename them. - `output_root` is where final Markdown, Excel, and Markdown assets are written. - Intermediate JSON and reports are written under `<output_root>/.summary-papers/`. - `output` controls final artifacts: - `markdown`: write one Markdown file. - `excel`: write one single-sheet `.xlsx` file. - `html`: write one static HTML reading page. - `all`: write Markdown, HTML, and Excel. ## Important environment rule Only claim tha