Today's Agent Skill: Inventory Demand Forecaster
What It Does
Running out of your best-selling products at the wrong moment costs sales and damages customer trust, while overstocking ties up cash and creates storage headaches. This skill analyzes your past sales history and flags what you should reorder — and when — before you hit a stockout.
How It Works
The agent takes a CSV or plain-text export of your past 90-180 days of sales by SKU or product name, calculates average weekly velocity and trend direction for each item, identifies your top movers, and produces a reorder priority list with suggested order quantities based on your lead times. It also flags items with accelerating sales velocity that may need larger buffers.
How to Deploy It
Run this skill in Claude Code, Gemini CLI, Cursor, or Windsurf by pasting or uploading a CSV export from your point-of-sale system (Square, Shopify, Clover, or QuickBooks), inventory platform, or even a manually maintained spreadsheet. Schedule a monthly run to refresh your reorder plan as seasonal patterns shift. Output can be pasted directly into a purchase order template or your inventory system's reorder form.
SKILL.md — Ready to Deploy
## Description
Analyzes 90-180 days of sales-by-SKU data to calculate weekly velocity, trend direction, and stockout risk for each product, then produces a ranked reorder list with recommended order quantities based on stated supplier lead times.
## Trigger
Invoke when the user says "forecast my inventory," "what do I need to reorder," "analyze my sales data for stockouts," or uploads a sales-by-SKU spreadsheet.
## Input
- Sales data: date, SKU or product name, units sold (CSV, TSV, or plain text table — 90 days minimum, 180 days preferred)
- Current on-hand quantity for each SKU (if known; mark as "unknown" if not)
- Average lead time in days from each supplier (or a single default if all suppliers have similar lead times)
- Safety stock preference: conservative (2 weeks buffer), standard (1 week buffer), or lean (3-day buffer)
- Optional: known upcoming demand spikes (holidays, seasonal events, planned promotions)
## Steps
1. Parse the sales data and compute average weekly units sold for each SKU over the full data window.
2. Split the data into two equal halves and compare average weekly velocity across the two periods. Classify each SKU as: Accelerating (>20% increase), StabCopy the full SKILL.md and drop it into your agent's skills directory to activate this skill.