Generate sourced PDF documents and answer questions about PDFs with page-level citations.
npx skills add https://github.com/b12io/b12-plugins --skill pdf-make-b12io
Please help me install this Agent Skill. Skill: pdf-make Source: https://github.com/b12io/b12-plugins/tree/main/plugins/pdf-make/skills/pdf-make Command: npx skills add https://github.com/b12io/b12-plugins --skill pdf-make-b12io
PDF p. 7, §9.2. - Measured print layout: Enforces explicit page size, margins, column widths, and built-in fonts so tables never silently run off the page edge. - Use Case: Paste your line items and ask for an invoice — you get a US Letter PDF with a ledger showing which facts came from you and which slots remain as brackets to fill. ## Quick Start Ask the assistant to turn your pasted notes or line items into a PDF invoice, or attach a PDF and ask what it says about a specific clause.High-intent search queries and answers about installing and using this skill.
Paste your line items, rates, and terms, then ask for the document type you need. The Skill builds a real .pdf file where every fact comes from your input, and anything missing — like a total or date — appears as a bracketed placeholder for you to fill.
Attach the PDF and ask your question directly. The Skill extracts text with pdfplumber, pypdf, or pdftotext, then answers with citations like "PDF p. 7, §9.2" and states exactly which pages it read.
Yes. Attach both versions and it produces a change table with one row per difference, citing the page and section in each version. Clauses it cannot locate on one side are reported as "not found" rather than assumed added or removed.
PDFs are built with reportlab, falling back to headless Chrome printing or a print-ready HTML file. Reading uses pdfplumber or pypdf first, then pdftotext -layout, then page rendering as an image if text extraction fails.
Auto-sized tables wider than the text frame overflow the page silently with no build error. This Skill prevents it by setting explicit column widths that sum to the text measure, and by rendering page one to visually check the result.
Do not use it for non-PDF prose like emails or blog posts, slide decks, .docx files, or image generation. Filling PDF forms, merging, and splitting are deferred to the built-in PDF skill instead.
anthropicsGenerate structured legal meeting briefings from calendars, emails, chats, and documents.
ZJU-REALIdentify contract risks and extract essential terms from TXT, Markdown, PDF, and DOCX documents.
HKU-MMLabAnalyzes tender pricing, spreadsheet totals, and vendor financial tradeoffs from local files.
phuhao00Create, modify, and analyze Microsoft Word .docx files programmatically.