/product-spec-pdf-parser

What it does

Type /as:product-spec-pdf-parser in Claude Code and point it at PDF files or a folder. The skill extracts text with the host’s native PDF tools, then Claude reads and reasons over the content to identify products, variants, SKUs, dimensions, materials, and pricing — structuring everything into reviewable rows with each value tied to its page, and unsupported values left unknown.

Handles price books (Herman Miller Aeron), fact sheets with SKUs (Hem Alphabeta), spec sheets, and multi-product catalogs. No regex or custom parsers — Claude IS the parser.

Install

Claude Desktop:

1. Open Customize → Browse plugins
2. Click + → Add marketplace from GitHub
3. Enter AlpacaLabsLLC/skills-for-architects
4. Install Architecture Studio

Claude Code (terminal):

claude plugin marketplace add AlpacaLabsLLC/skills-for-architects
claude plugin install as@skills-for-architects

Or symlink just this skill:

git clone https://github.com/AlpacaLabsLLC/skills-for-architects.git
ln -s $(pwd)/skills-for-architects/skills/product-spec-pdf-parser ~/.claude/skills/product-spec-pdf-parser

Usage

/as:product-spec-pdf-parser

Then provide PDF paths or a folder:

/as:product-spec-pdf-parser

~/Documents/specs/alphabeta-floor-lamp.pdf
~/Documents/specs/aeron-price-book.pdf

Choose variant depth — expand (one row per variant, default) or summarize (one row per product).

Output schema

Fields like these per product. Library saves use the 33-column product schema, with variant, price adder and origin kept in Notes:

Field	Example
Product Name	Alphabeta Floor Lamp
Variant	Diamond, Black
SKU	HEM-AF-DB
Brand / Designer	Hem / Luca Nichetto
Collection / Category	Alphabeta / Lighting
Description	Modular floor lamp with interchangeable shades
W / D / H / Seat H	— / — / 135.5 / —
Unit / Weight	cm / 4.5 kg
Materials	Aluminium, Steel
Colors/Finishes	Black
List Price / Price Adder	595.00 / —
Currency	EUR
Warranty / Certifications	— / CE
Country of Origin	Sweden
Source File	alphabeta-fact-sheet.pdf

PDF types supported

Type	Strategy
Fact sheet with SKUs	One row per SKU
Fact sheet with finishes	One row per upholstery/finish option
Price book / configurator	One row per product type, options summarized
Product catalog	Rows for each distinct product

How it handles failures

• Scanned/image PDFs — detected and flagged for OCR
• Password-protected PDFs — caught and reported
• Large PDFs (100+ pages) — processed in bounded chunks, with per-page coverage and gaps reported

After every batch: Parsed: X products from Y PDF(s)

Pairs with

Use /as:product-spec-bulk-fetch for web URLs instead of PDFs, then /as:product-data-cleanup to normalize. Chain: PDF parse → cleanup → spec-ready schedule.