assets/skills/claude-design/read-pdf/SKILL.md

---
name: read-pdf
description: Use when extracting text, tables, or content from PDF documents
---

# Read PDF

To read a PDF in run_script, use the browser build of pdf-parse (pinned @2.4.5):

```js
const { PDFParse } = await import('https://cdn.jsdelivr.net/npm/pdf-parse@2.4.5/dist/pdf-parse/web/pdf-parse.es.js');
PDFParse.setWorker('https://cdn.jsdelivr.net/npm/pdf-parse@2.4.5/dist/pdf-parse/web/pdf.worker.min.mjs');

const blob = await readFileBinary('document.pdf');
const parser = new PDFParse({ data: new Uint8Array(await blob.arrayBuffer()) });
const result = await parser.getText();
log(result.text);
```

SRI hashes (for reference — dynamic import() cannot enforce SRI at runtime):
- `pdf-parse.es.js` — sha384-J7LMAGioDDEBxHBcdxpU9NGtQu2/iLuSGyD3HsO5aYDJ0BAisPtpTYGc5XcB7UcI
- `pdf.worker.min.mjs` — sha384-zdw/VQhL/JrSgvr/Omai4B8USJUC6AQXr/4YW01OlVWutKoGvg34AOFCRsO1dGJr