agentsclimarketplace
Sign in

Paper figure extractor

Skill StaryMoon/paper-figure-extractor-skill/skills/paper-figure-extractor

Extract figures, captions, page references, contact sheets, and a machine-readable manifest from research PDFs. Use when Codex needs to collect paper visuals for literature notes, presentations, reproduction work, figure comparison, or citation-aware review without manually taking screenshots.From its SKILL.md

Install
npx -y skills add StaryMoon/paper-figure-extractor-skill --skill paper-figure-extractor

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

1.9 KB, 326 tokens by cl100k_base, as published. Nobody here has run it

Paper Figure Extractor

Create a traceable figure bundle from a research PDF. Keep every extracted image linked to its source page and preserve captions as evidence rather than guessing missing context.

Workflow

  1. Check that the input is a readable PDF and create a new output directory.
  2. Run the extractor:
python3 scripts/extract_paper_figures.py paper.pdf --output output/paper-figures
  1. Inspect contact-sheet.png before using the assets. Remove publisher logos, equation fragments, or decorative images only after visual review.
  2. Read manifest.json when downstream work needs page numbers, dimensions, hashes, or caption candidates.
  3. Use report.md as the human-facing index. Keep the source PDF path and page numbers when moving assets elsewhere.

Quality rules

  • Never claim an extracted bitmap is a complete figure unless the contact sheet confirms it.
  • Treat captions as candidates when the PDF text layer is malformed.
  • Keep small images only when the user explicitly asks for icons, equations, or logos.
  • Do not silently upscale source assets.
  • If the PDF contains only vector drawings, render the relevant page and record that fallback in the report.

Options

  • --min-width and --min-height: filter tiny embedded assets.
  • --include-small: retain every embedded image.
  • --pages 1,3-5: process selected one-based pages.
  • --overwrite: replace an existing output directory.

See references/output-contract.md before changing filenames or manifest fields.

What ships with it: 3 files

7.2 KB alongside SKILL.md, 1 of them executable

agents/

references/

scripts/

Keep looking

Skills are one crate of 326,144. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.