# Supply Chain Prover MCP

> The Supply Chain Prover forces five critical supply chain axes into a single validation check: demand forecasting, inventory optimization, supplier risk, logistics costs, and bullwhip effect mitigation. It tells you if your plan is based on math or gut feeling. Stops catastrophic failures before they happen.

## Overview
- **Category:** operations
- **Price:** Free
- **Tags:** supply-chain, inventory, logistics, toyota, eoq, bullwhip-effect, demand-forecasting

## Description

You’re running numbers on a spreadsheet for hours, but that math might miss half the risks. The `validate_supply_chain` tool forces your entire operation through five critical supply chain checks simultaneously: demand forecasting, inventory optimization, supplier risk, logistics costs, and bullwhip effect mitigation.

The **Supply Chain Prover** doesn't tell you if your plan is good; it tells you if it’s based on solid math or just a gut feeling. It stops catastrophic failures before they even happen by making sure every assumption holds up under rigorous validation.

When you call `validate_supply_chain`, the system immediately runs five checks: **Demand Forecasting** validates your historical sales data using statistical models to provide a measurable forecast complete with confidence intervals and Mean Absolute Percentage Error (MAPE) metrics. For inventory, it calculates **Optimal Inventory Levels**, determining both the Economic Order Quantity (EOQ) and necessary safety stock for every single SKU based on carrying costs and demand variability math.

The tool also audits your supplier base to check **Supplier Concentration Risk**, verifying geographic spread and ensuring that no single source accounts for more than 30% of any critical component category. It analyzes **Logistics Cost Per Unit** by comparing the total cost per unit across different transport modes—air, sea, or road—and figuring out the true last-mile shipping percentage. Finally, it measures your **Bullwhip Effect Potential**, assessing how demand spikes might amplify up through your supply chain tiers by checking both Point-of-Sale (POS) data sharing frequency and pricing stability.

This isn't a dashboard that spits out pretty graphs; it’s an auditor that forces mathematical proof for every decision you make.

## Tools

### validate_supply_chain
Calls five checks simultaneously: demand forecasting (MAPE), inventory math (EOQ, safety stock), supplier risk percentage, logistics cost-per-unit analysis, and bullwhip effect mitigation.

## Prompt Examples

**Prompt:** 
```
We expect demand to grow next quarter. We keep enough stock from our trusted supplier. Logistics handles shipping. Orders follow demand.
```

**Response:** 
```
GUT_FEEL_FORECAST — Five failures: no statistical model, no inventory math, single supplier, logistics handwaving, bullwhip amplification.
```

**Prompt:** 
```
Demand: exponential smoothing (α=0.3) on 24 months, forecast 12,400 ± 1,800 units/month (95% CI), MAPE 11.2%. Inventory: EOQ 1,467 units, safety stock 1,005, reorder point 6,797, carrying 24.8%. Suppliers: 3 suppliers — 38% Vietnam, 35% India, 27% Mexico. Lead time CV 0.22. Logistics: 60% sea ($0.18/kg), 30% road ($1.20/kg), 10% air urgent, cost/unit $2.34, last-mile 47%. Bullwhip: daily POS EDI to top-3 suppliers, weekly orders (not monthly), EDLP reduced forward-buying 35%, lead time compressed 45→28 days.
```

**Response:** 
```
SUPPLY_CHAIN_PROVEN — All five axes validated at Toyota level.
```

**Prompt:** 
```
Demand: ARIMA(1,1,1) on 36 months, 8,200 units/month ± 900 (95% CI), MAPE 8.7%. Inventory: EOQ 982 units, safety stock 645, reorder 4,120, carrying 21.3%. Suppliers: 2 suppliers — 72% Guangdong, 28% Guangdong (same region). Lead time CV 0.41. Logistics: 85% sea, 15% air, cost/unit $1.90. Bullwhip: monthly order batches.
```

**Response:** 
```
SINGLE_SOURCE_NAIVE — Forecast and inventory pass. Suppliers FAIL: 100% in Guangdong = catastrophic geographic concentration. Lead time CV 0.41 = unreliable. Diversify to at least 2 regions. Then fix monthly batching (bullwhip amplification).
```

## Capabilities

### Validate Demand Forecasting
Runs statistical models on historical sales data to provide a measurable forecast with confidence intervals and MAPE metrics.

### Calculate Optimal Inventory Levels
Determines the Economic Order Quantity (EOQ) and necessary safety stock based on carrying cost and demand variability math for every SKU.

### Audit Supplier Concentration Risk
Checks your supplier base to ensure geographic spread and limits single-source reliance below 30% of any critical component category.

### Analyze Logistics Cost Per Unit
Compares the total cost per unit using different transport modes (air, sea, road) and determines last-mile shipping percentage.

### Measure Bullwhip Effect Potential
Assesses how demand spikes might amplify through your supply chain tiers by checking POS data sharing frequency and pricing stability.

## Use Cases

### Post-Disruption Planning
A company just recovered from a chip shortage. They ask their agent to run `validate_supply_chain` using the new supplier base data. The tool flags that while they diversified, their lead time coefficient of variation (CV) is still too high, forcing them to renegotiate stricter delivery reliability metrics.

### New Product Launch Strategy
The product team wants to launch a new line. Instead of ordering on gut feel, they run `validate_supply_chain`. The tool immediately flags that their current raw material supplier is 100% in one region (Shenzhen), forcing the team to implement a dual-source plan before production even begins.

### Optimizing Basic Goods Shipments
A clothing brand needs to ship 200,000 basic items. They ask their agent to run `validate_supply_chain`. The tool's logistics check shows that air freight is overkill for these basics, calculating that switching 90% of the volume to sea freight saves hundreds of thousands in annual costs.

### Correcting Inventory Overstock
A restaurant chain keeps ordering napkins 'when they run low.' They use `validate_supply_chain`. The tool calculates the optimal EOQ, revealing that their current over-ordering strategy is costing them tens of thousands annually in unnecessary delivery and holding fees.

## Benefits

- **Stop Inventory Blindness:** Instead of guessing how many napkins to buy, `validate_supply_chain` calculates the precise Economic Order Quantity (EOQ) and required safety stock for every SKU, slashing excess ordering costs.
- **De-Risk Your Sourcing:** The tool forces you to audit supplier concentration. It flags if one factory holds too much risk (over 30%) in a critical category, preventing catastrophic single-source failures.
- **Cut Logistics Waste:** You stop shipping everything by air freight because of 'urgency.' The Prover compares mode costs (sea vs. road) and calculates the true cost/unit shipped for basic goods versus trend items.
- **Fix Bad Forecasting:** Forget 'we expect growth.' This tool demands statistical proof, requiring a defined model, historical data period, and a MAPE target under 15% to make your forecast actionable.
- **Prevent Bullwhip Amplification:** By checking POS data sharing across tiers, the Prover stops inflated order signals from causing massive overstocking and inventory write-downs months later.

## How It Works

The bottom line is: you move from gut-feeling business plans to mathematically defensible operations.

1. Feed the tool all relevant operational data: 24+ months of historical sales, current inventory levels (SKU-by-SKU), supplier contracts, and recent logistics costs.
2. The `validate_supply_chain` tool runs five separate mathematical checks—forecasting, EOQ calculation, risk audit, cost analysis, and bullwhip assessment—against your input data.
3. You get back a structured report. It doesn't just say 'fail'; it points to the exact math that is wrong (e.g., 'EOQ needs adjustment,' or 'Supplier concentration exceeds 30%').

## Frequently Asked Questions

**What is the difference between a forecast and what Supply Chain Prover does?**
A standard forecast predicts ('We expect growth'). The Prover validates that prediction using statistical proof. It demands historical MAPE metrics, a defined model (like ARIMA), and a confidence interval to be actionable.

**How does Supply Chain Prover handle single-supplier risk?**
It audits supplier concentration percentage. If one supplier makes more than 30% of any critical component, it flags the risk immediately, forcing you to build a dual-source plan.

**Can I use Supply Chain Prover for anything other than manufacturing?**
Yes. The tools are universal. Whether you're dealing with electronics or consumer goods, it checks the underlying math: EOQ, risk diversification, and logistics cost-per-unit.

**Do I need to share POS data for Supply Chain Prover?**
Yes, that’s key. The tool requires Point-of-Sale (POS) data sharing across tiers to properly assess the bullwhip effect and prevent massive overordering.

**What specific data metrics does Supply Chain Prover require for inventory optimization?**
It needs cost per unit, annual demand (D), ordering cost (S), and carrying cost (H) for every SKU to calculate the optimal EOQ. For safety stock, you must provide service level inputs, demand variability ($\sigma$), and accurate lead time (L). These metrics are foundational; without them, the calculations fail.

**How does Supply Chain Prover assess supplier risk across multiple geographic regions?**
The tool checks for both geographic spread and concentration percentage per region. It mandates dual-sourcing for any component with a lead time over two weeks, preventing reliance on single points of failure. You must source from at least two different global regions.

**What is the minimum required data cadence for Supply Chain Prover to remain accurate?**
You need high-frequency data feeds—daily POS or inventory updates are ideal. The system requires minimizing order batching, meaning you should plan for smaller, more frequent orders rather than relying on monthly reporting cycles.

**How does Supply Chain Prover prevent gut-feel errors in planning?**
It demands a statistical foundation for all predictions. Any demand forecast must include the model name (e.g., ARIMA), a minimum data period of 24 months, and calculate a Mean Absolute Percentage Error (MAPE) score that is under 15% to be considered actionable.

**Why is gut-feel forecasting dangerous?**
'We expect demand to grow' is a hope, not a forecast. Statistical forecasting uses historical decomposition (trend + seasonality + noise), selects the right model (exponential smoothing for stable demand, ARIMA for complex patterns), provides confidence intervals ('10,000 ± 1,500 units at 95%'), and measures accuracy with MAPE. Toyota measures forecast error weekly. What is yours?

**What is the bullwhip effect?**
A 10% demand increase at retail becomes a 20% order increase at distributor, 40% at manufacturer, and 80% at raw material supplier. Each echelon amplifies the signal 2-5x. Causes: order batching, price fluctuations, demand forecasting errors, lead time inflation. Mitigate with: POS data sharing upstream (Walmart/P&G model), smaller and more frequent orders, price stabilization, and lead time compression.

**Why does last-mile cost 40-53% of total logistics?**
Container shipping moves 20,000 TEUs at $0.10-0.30/kg. A truck moves 20 tons at $0.50-2.00/kg. A delivery van moves 200 packages at $5-15 each. Each step loses economies of scale. The last mile has the smallest vehicles, most stops, most failed deliveries (15-20% not-at-home), and highest labor cost per unit. Solving last-mile is a $100B+ industry problem.