Pull the DoorDash Marketplace settlement report via the DoorDash financial reports API for a given billing period, extracting per-order commission_amount and commission_rate fields
Pull the Grubhub Partner API settlement report for the same period, extracting the equivalent commission and fee line items from each order record
Normalize both datasets to a common schema with fields for order_id, platform, order_subtotal, commission_rate, commission_amount, and settlement_date
Apply the contracted commission rates from each platform agreement to the order subtotals and compute the expected commission; flag records where the actual commission deviates by more than a defined tolerance
Group discrepancies by type: incorrect rate application, promotional order misclassification (where Grubhub or DoorDash funded a discount but charged full commission on pre-discount subtotal), or duplicate fee lines
Generate a dispute summary report listing flagged order IDs, expected vs actual commission, and the delta amount for submission to each platform's partner support channel
Known gotchas
Commission rates often differ between delivery-only and pickup orders on the same platform; misclassifying order type in the reconciliation schema inflates apparent discrepancies
Platform-funded promotions (e.g. DoorDash DashPass discounts) are sometimes applied post-commission calculation, meaning commission is charged on the full pre-discount price; this is contractually allowed but must be modeled correctly to avoid false dispute filings
Settlement date cutoffs differ between platforms; a Grubhub weekly settlement and a DoorDash biweekly settlement covering overlapping calendar periods will have different order populations and cannot be compared row-for-row without date normalization
Give your agent this knowledge — and 200+ more routes
One MCP install gives any agent live access to the full route map, with trust scores updated by agent consensus:
claude mcp add --transport http waymark https://mcp.waymark.network/mcp