Purchase Order Processing

Purchase Order Processing Skill

When to Use

Activate this skill when the user wants to:

• Process a purchase order (PO) and generate a sales order
• Match purchase order items with bills of material (BOM)
• Calculate pricing from a price list
• Generate a sales order document from PO data
• Extract and enrich order data from customer documents

Trigger phrases:

• "process purchase order", "process PO", "process this order"
• "generate sales order from PO"
• "calculate pricing for this order"
• "match BOM with purchase order"
• "create sales order"

Inputs

Required Folder Structure

The user provides a folder path containing:

input_folder/
├── customer_data/           # REQUIRED: Customer documents
│   ├── PO.pdf              # Purchase Order (required)
│   ├── BOM1.pdf            # Bill of Materials (optional)
│   ├── BOM2.pdf            # Additional BOMs as needed
│   └── ...
├── price_list/              # REQUIRED: Pricing information
│   └── price_list.xlsx     # Master price list (or .csv)
│
└── sample_sales_order/      # OPTIONAL: Sample for format inference
    └── sample_order.docx   #   (if not provided, use default text format)

Input Parameters

Tooling Rules (Windows vs Linux Path Safety)

Why this matters

On Windows, Claude Desktop + toolchains sometimes behave like they are in a POSIX shell, producing paths like /mnt/c/.... Meanwhile, your MCP servers may run native Windows Python, expecting C:\.... This mismatch can cause "file not found" or failing shell commands.

Strict rules

1. Prefer MCP filesystem tools for file/folder operations Use the filesystem server for listing and reading files instead of shell commands.

2. Avoid bash commands on Windows If you must run a command on Windows, prefer PowerShell.

3. Never assume the environment is Linux Treat the runtime as OS-ambiguous and enforce the above rules to stay stable.

4. Never do the following: NEVER run pip install, python -c, pdfplumber, or any ad-hoc parsing code for .pdf/.pptx/.xlsx.

NEVER use /mnt/user-data/uploads/... paths; only use paths returned by the MCP filesystem listing or the user-provided Windows folder.

If you are about to do any of the above, STOP and switch to the built-in PDF/PPTX/XLSX skills.

• View /mnt/skills/public/docx/SKILL.md skill for reading and editing DOCX files.
• View /mnt/skills/public/pdf/SKILL.md skill for reading PDF files.
• View /mnt/skills/public/xlsx/SKILL.md skill for reading and editing XLSX files.

Workflow

Execute these steps in order. Reference the files in reference/ folder for field definitions and customization options.

Step 0: Validate Input Folder (MCP filesystem only)

1. Verify the input folder exists
2. Check for required subfolders:
   • customer_data/ - Must exist and contain at least one PDF (the PO)
   • price_list/ - Must exist and contain price list file
3. Identify optional folders:
   • sample_sales_order/ - Check if sample document exists
4. If required folders/files missing, ask user to provide them

Output: Inventory of available files

If customer_data/ is missing or empty, stop and ask the user to add the source materials there.

Step 1: Extract Fields from Purchase Order

Read the Purchase Order PDF from customer_data/ and extract all fields as defined in reference/EXTRACTION_FIELDS.md.

Extract:

• PO Header: Order number, date, payment terms, shipping terms, project code
• Customer Information: Name, address, contact details
• Line Items: Material numbers, descriptions, quantities, units, delivery dates
• PO Summary: Subtotal, shipping, tax, total
• Special Requirements: Instructions and documentation requirements

See reference/EXTRACTION_FIELDS.md for complete field specifications, source locations, and required/optional status.

Output: Structured PO data with all extracted fields

Step 2: Extract Fields from Bills of Material

For EACH Material_Number found in the PO line items:

1. Search customer_data/ for BOM files
2. Find the BOM where BOM_ID matches Material_Number
3. If match found, extract all fields as defined in reference/EXTRACTION_FIELDS.md

Extract:

• BOM Header: Revision, date, status, customer PO reference, project code
• Material Identification: BOM ID, Type/Part Designation (critical for price lookup), dimensions, material grade
• Technical Specifications: Material standard, surface finish, weight, testing requirements
• Fee Calculation Fields: Cutting_Required, Testing_Lots_Required, Certificates_Required (for determining applicable fees)

See reference/EXTRACTION_FIELDS.md for complete field specifications, source locations, and required/optional status.

Decision Point:

• If BOM found for all items → Continue to Step 3
• If some BOMs missing → Check if PO contains the Type/Part Designation info
• If PO has all info (no BOMs needed) → Mark as "PO-only mode" and continue
• If critical info missing → Flag items for manual review

Output: BOM data for each Material_Number, or flags for missing BOMs

Step 3: Match PO Items with BOMs

Create enriched items by combining PO and BOM data:

FOR EACH PO Line Item:
    MATCH WHERE: PO.Material_Number = BOM.BOM_ID

    IF match found:
        CREATE Enriched_Item with:
        - Item_Number (from PO)
        - Material_Number (from PO, validated against BOM)
        - Type_Part_Designation (from BOM) ← CRITICAL for price lookup
        - Dimensions (from BOM)
        - Material_Grade (from BOM)
        - Quantity (from PO)
        - Unit (from PO)
        - Delivery_Date (from PO)
        - Weight_Per_Unit (from BOM)

    ELSE IF PO contains Type/Part info:
        CREATE Enriched_Item using PO data only

    ELSE:
        FLAG as "BOM Not Found - Manual Review Required"

Validation Rules

Output: List of Enriched_Items ready for pricing

Step 4: Lookup Prices from Master Price List

Read the price list from price_list/ folder and match each enriched item using the matching logic defined in reference/PRICING_RULES.md.

For EACH Enriched_Item:

1. Match using Type/Part Designation (primary matching key). There should be an exact match.
2. If match found, extract unit price and fees (cutting, testing, certification)
3. If no match found, flag item as "Price Not Found - Manual Pricing Required"

See reference/PRICING_RULES.md for:

• Price list column mapping
• Complete matching logic
• Fallback matching rules

IMPORTANT: Never use the PO's unit prices. Always use the Price List prices.

Output: Priced_Items with all pricing fields assigned

Step 5: Calculate Pricing

Apply calculations as defined in reference/CALCULATION_LOGIC.md and reference/PRICING_RULES.md.

5.1 Line Item Calculations

For EACH Priced_Item, calculate:

• Material_Cost = Quantity × Unit_Price
• Cutting_Cost (if applicable)
• Testing_Cost
• Cert_Cost
• Line_Total

See reference/CALCULATION_LOGIC.md for detailed formulas, defaults, and examples.

5.2 Material Subtotal

Calculate the sum of all Material_Cost values across all items.

See reference/CALCULATION_LOGIC.md for calculation details.

5.3 Volume Discount

Determine the applicable discount tier based on the material subtotal amount and apply the discount rate.

See reference/PRICING_RULES.md for complete discount tier table and rules.

5.4 Fees, Shipping, Tax, and Grand Total

Aggregate all fees (cutting, testing, certification), add shipping from PO, calculate tax on the taxable base, and compute the grand total.

See reference/CALCULATION_LOGIC.md for detailed formulas and tax rules.

Output: Complete pricing breakdown for all items and order totals

Step 6: Generate Sales Order

Create the Sales Order document using format from reference/OUTPUT_FORMAT.md.

6.1 Determine Output Format

• IF sample_sales_order/sample_order.docx exists:
  1. Use DOCX skill to READ the sample document
  2. Extract the EXACT pricing summary structure:
     • Field labels (verbatim)
     • Field order (must be preserved)
     • Currency formatting rules
     • Punctuation (colons, parentheses)
  3. Use this as the TEMPLATE for output
  4. MATCH every detail: labels, order, formatting
• ELSE → Use default format from reference/OUTPUT_FORMAT.md