Turn sales history and current stock counts into reorder points, order quantities, and a this-week purchase list -- so a product business stops stocking out of winners and stops burying cash in losers. Handles seasonality, supplier lead times, and the dead-stock cleanup.
SkillScore70Maintained
Automatic
◆Procurement pack
Last verified 8/24/2026
What it changes
Passed 16 of 23 test cases with the skill, and 10 of 23 without it.
Without skill43%
With skill70%
Checks it passed that the model alone missed
Refocus on stock replenishment planning.
Classify items into ABC tiers based on revenue contribution.
Designate items with zero sales history as NEW -- judgment order.
and 4 more in the full report
6 fixed10 passed either way7 still failing0 broke
Pass-rate lift
+26 pp
Tokens
+14%
Agent turns
0%
Response time
-28%
Model gemini-3.6-flashMeasured 8/24/2026v116 of 23 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 6 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 6
Prompt
Here are our calculated order quantities for Supplier X: SKU-1: 14.2 units, SKU-2: 7 units. Can we send these raw quantities directly on our weekly purchase orders?
What was checked, without and with the skill
Checked
Without skill
With skill
Request supplier constraints such as minimum order quantities or case packs.
Missed
Passed
Show explicit rounding adjustments when meeting supplier order constraints.
Passed
Passed
case-09 · gemini-3.6-flash · 8/24/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md27 linessingle file
Inventory Reorder Planner
Stockouts burn revenue on the products that sell; overstock buries cash in the ones that do not. Both come from ordering by gut. Input: sales history export (POS, Shopify, marketplace, or spreadsheet), current stock counts, and supplier lead times (asked per supplier if not provided). Output: reorder points per SKU and the purchase list for this week.
Workflow
Profile every SKU. From sales history: average daily/weekly velocity, variability, trend (growing/flat/dying), and seasonality where the history is long enough to show it (holiday spikes, summer lulls) -- flagged per SKU, never assumed.
Classify. ABC by revenue contribution (A: the ~20% of SKUs driving ~80%), crossed with velocity: WINNERS (high velocity, healthy margin -- never stock out), STEADY, SLOW, DEAD (no sales in 90+ days with stock on hand).
Compute reorder points. Per SKU: lead-time demand (velocity x supplier lead time) plus safety stock scaled to demand variability and the SKU's class -- A items get generous buffers, C items get lean ones. Show the formula inputs per SKU so the owner can sanity-check against what they know ("velocity 4/wk, lead time 3 wks, safety 1.5 wks = reorder at 18").
The purchase list.reorder-plan.csv: every SKU at or below reorder point, recommended order quantity (covering the review cycle plus lead time, respecting supplier minimums and case packs when known), cost, and supplier -- grouped by supplier to build actual POs. Flag URGENT where projected stockout lands before the order can arrive.
Free the cash. The dead-stock report: units, cost tied up, days since last sale, and the exit per item -- bundle with a winner, discount, donate for the writeoff, or return to supplier. Cash sitting in dead stock is the cheapest funding most product businesses never tap.
Rules
Velocity from real sales only; no history means NEW -- judgment order, clearly separated from computed recommendations.
Stockout periods bias velocity down -- when the data shows zero-stock stretches, compute velocity from in-stock days only and note it.
Seasonality is applied only where the data shows it, and the applied multiplier is stated per SKU.
Supplier constraints (minimums, case packs, price breaks) are asked for, and rounding to meet them is shown, not silent.
A items err toward buffer, C items err toward lean -- one safety-stock policy for everything is the amateur tell.
Re-run on a cadence (weekly or per supplier order cycle); each run compares last cycle's recommendation to what actually sold and tightens.
Quick Commands
"Plan from [sales export] and [stock counts]" -- full workflow
"What do I order this week?" -- the purchase list only