desktop

Navigate

Overview DesktopBench Leaderboard RL Envs Evaluation
Main site Evals Blog Dashboard
core candidate

StoreOps-200

Commerce operator workflows across product catalog, inventory, orders, refunds, promotions, shipping, buyer chat, and sales analytics.

Core release candidate: StoreOps-50

Task families

20 families x 10 variants

ProductCreateProductVariantProductEditProductPublishInventoryAdjustInventoryTransferOrderFulfillOrderEditReturnApproveRefundPartialCancelRejectVoucherCreateDiscountCreateBundleDealAddOnDealShippingRateStoreAnnouncementBuyerChatReplySalesAnalyticsStoreMode
Target UIs

Mock apps for resettable rollouts

Shopify-like admin

SmartStore-like admin

Coupang Wing-like seller center

Cafe24-like storefront

Examples

Representative task contracts

Example 1

StoreOps.OrderFulfill.001

Instruction
Fulfill order #1048 using tracking number CJ123456789KR and carrier CJ Logistics.
Grader
order.status == fulfilled, carrier == CJ Logistics, tracking_number == CJ123456789KR
Example 2

StoreOps.ProductEdit.001

Instruction
Find Acrylic Phone Stand and reduce the price from KRW 12,000 to KRW 9,900.
Grader
product.price == 9900 and unrelated product fields remain unchanged
Example 3

StoreOps.BuyerChatOrder.001

Instruction
Buyer sora12 asks where order #1071 is. Check the order status and reply with the tracking number.
Grader
correct chat thread contains current tracking number from order #1071
Verifier

Suite-level verifier shape

DB state checks for product, order, inventory, promotion, and message records plus side-effect guards.

On this page

  • Summary
  • Task families
  • Examples
  • Verifier

Help improve UseDesktop

Optional analytics help us learn which pages are useful. They stay off unless you accept.

Privacy