R05.08.01.02 · TASK CLASS

Orders, returns & marketplace operations

Look up and report on orders and sales, and process orders, refunds and marketplace listings within policy.

20 project-authored proposals · taxonomy v1.1 · 0 new benchmark runs. Verifiers and partial-credit rules below are designs, not executed results.

Typical applications: Shopify admin · Amazon Seller Central
Class GUI-need label: HY · Hybrid

Task proposals · 20

Expand to inspect the complete original wording and proposed checks. Original proposal text remains English in both interfaces.

R05.08.01.02-T01 Refund a damaged line item without restocking

Original instruction

Shopify development store 'Northwind Outdoor' is open in the admin. Customer email fixtures/email_1043.eml reports one 'Trail Mug - Blue' from order #1043 arrived chipped. Per input/returns_policy.md, refund the item plus its tax, do not restock damaged goods, do not refund shipping and do not notify the customer.

Platforms: web · Any OS (web)
Applications: Shopify admin
Interfaces: brow · Browser
GUI need · task setting: HY · Hybrid
GUI need · ordinary access: G2 · GUI-essential
Interaction primitives: click · Point & click type · Text & forms multi · Cross-app
Work product: WP08 · App/system state
Conditions: None declared
Proposed difficulty: 2 / 5

Proposed verifier

Admin API read-back: exactly one new refund on #1043 with one refundLineItem (qty 1, restockType NO_RESTOCK), amount 21.60, shipping refund 0, notify=false; inventory levels unchanged. Agent gets no API token; order timeline actor must be staff user 'ops_agent', not an app.

Proposed partial credit

40% correct line and qty; 25% amount incl. tax, no shipping; 25% no restock; 10% no notification and no other order changes.

R05.08.01.02-T02 Fulfill priority orders across two locations with tracking

Original instruction

In the Shopify admin of development store 'Northwind Outdoor' (test gateway), fulfill the 6 unfulfilled orders tagged 'priority'. Ship each line from the location that holds stock; #1088 must be split between Brooklyn and Reno. Enter UPS tracking numbers from fixtures/tracking.csv. Leave any order with a High fraud-risk recommendation unfulfilled and tag it 'hold-risk'.

Platforms: web · Any OS (web)
Applications: Shopify admin
Interfaces: brow · Browser
GUI need · task setting: HY · Hybrid
GUI need · ordinary access: G2 · GUI-essential
Interaction primitives: click · Point & click type · Text & forms multi · Cross-app
Work product: WP08 · App/system state
Conditions: None declared
Proposed difficulty: 3 / 5

Proposed verifier

GraphQL read-back of fulfillments: per order the location id, line quantities and tracking number/carrier match key; #1088 has 2 fulfillments; high-risk #1091 has 0 fulfillments and tag 'hold-risk'. Timeline actor checked to block API-app shortcuts.

Proposed partial credit

40% 5 orders fulfilled with exact tracking; 25% correct split on #1088; 25% #1091 held and tagged; 10% no customer-facing extra notes.

R05.08.01.02-T03 Process a size exchange through Shopify returns

Original instruction

Order #1120 (Shopify development store 'Northwind Outdoor', test gateway) has a return request: customer wants 'Ridge Fleece' size M exchanged for size L. Approve the return, create the exchange line for size L from Reno stock, apply no price difference, and mark the returned M as received and restocked at Reno once the return is created.

Platforms: web · Any OS (web)
Applications: Shopify admin
Interfaces: brow · Browser
GUI need · task setting: HY · Hybrid
GUI need · ordinary access: G2 · GUI-essential
Interaction primitives: click · Point & click type · Text & forms
Work product: WP08 · App/system state
Conditions: None declared
Proposed difficulty: 3 / 5

Proposed verifier

Read Return object for #1120: status CLOSED or OPEN with returnLineItem M qty 1 received and restocked at Reno; exchangeLineItem variant L qty 1; balance due 0.00; Reno inventory M +1, L reserved -1. Timeline actor must be staff user.

Proposed partial credit

30% return approved on correct line; 30% exchange variant L; 25% received and restocked at Reno; 15% zero balance, no refund issued.

R05.08.01.02-T04 Triage pending return requests in Seller Central by policy

Original instruction

A sandbox replica of Amazon Seller Central is open on Manage Returns with 9 pending return requests. Apply input/return_rules.md: authorize requests within 30 days of delivery, close out-of-window requests with reason 'Outside return window', and for 'item defective' reasons over $50 issue a returnless refund. Record nothing else.

Platforms: web · Any OS (web)
Applications: Amazon Seller Central
Interfaces: brow · Browser
GUI need · task setting: G3 · GUI-only
GUI need · ordinary access: G2 · GUI-essential
Interaction primitives: click · Point & click nav · Pan/zoom/scroll type · Text & forms
Work product: WP08 · App/system state
Conditions: None declared
Proposed difficulty: 3 / 5

Proposed verifier

Replica DB: each of 9 RMA rows has status/reason/refund flag matching the rule key computed from delivery dates and order values. The replica issues no SP-API/LWA credentials and exposes no API; it accepts writes only from logged-in browser sessions, checked against the UI event log.

Proposed partial credit

90% requests decided per rule (pro rata over the 9); 10% no extra refunds or messages.

R05.08.01.02-T05 Answer three A-to-z claims using carrier evidence

Original instruction

In the Seller Central replica, Performance > A-to-z Guarantee Claims lists 3 open 'item not received' claims. For each, open the tracking in the order page and fixtures/carrier_scans.pdf. Per input/claims_policy.md, respond with proof (carrier, tracking, delivery date, ZIP) when delivered to the ship-to ZIP with signature; otherwise issue a full refund from the claim page.

Platforms: web · Any OS (web)
Applications: Amazon Seller Central
Interfaces: brow · Browser
GUI need · task setting: G3 · GUI-only
GUI need · ordinary access: G2 · GUI-essential
Interaction primitives: click · Point & click type · Text & forms multi · Cross-app nav · Pan/zoom/scroll
Work product: WP08 · App/system state
Conditions: None declared
Proposed difficulty: 4 / 5

Proposed verifier

Replica claim table: per claim action (respond/refund) vs key; response text must contain the exact tracking number, carrier and delivery date (regex); refunds equal order total. The replica issues no SP-API/LWA credentials and exposes no API; it accepts writes only from logged-in browser sessions, checked against the UI event log.

Proposed partial credit

90% claims decided correctly with complete evidence text (one third per claim); 10% no duplicate refunds.

R05.08.01.02-T06 Reinstate suppressed listings with a compliant main image

Original instruction

Seller Central replica, Inventory > Fix Your Products shows 4 suppressed listings. For each, choose from assets/<SKU>/ the one candidate image that meets the main-image rule (product only, pure white background, fills >=85% of frame), set it as MAIN, and add any missing bullet or brand field from fixtures/catalog.xlsx.

Platforms: web · Any OS (web)
Applications: Amazon Seller Central · Excel
Interfaces: brow · Browser
GUI need · task setting: G2 · GUI-essential
GUI need · ordinary access: G2 · GUI-essential
Interaction primitives: click · Point & click vis · Visual judgment type · Text & forms multi · Cross-app
Work product: WP08 · App/system state
Conditions: None declared
Proposed difficulty: 4 / 5

Proposed verifier

Replica listing rows: MAIN image hash equals the compliant candidate (key built by measured background RGB=255 and bbox fill); brand/bullets match catalog; suppression flag cleared for all 4. The replica has no feed endpoint and issues no API token; images are accepted only via the browser image manager, matched to the UI event log.

Proposed partial credit

60% correct main images (one quarter per SKU); 30% missing fields filled exactly; 10% all 4 active.

R05.08.01.02-T07 Push a guarded price update through an inventory file upload

Original instruction

Merchandising sent fixtures/price_changes.xlsx (64 SKUs, new price, floor, ceiling). In the Seller Central replica, build the price-and-quantity inventory file from the template under Add Products via Upload, upload it, read the processing report, fix the rejected rows (e.g. price below floor, bad SKU case) and re-upload until 0 errors.

Platforms: web · Any OS (web)
Applications: Amazon Seller Central · Excel
Interfaces: brow · Browser
GUI need · task setting: HY · Hybrid
GUI need · ordinary access: HY · Hybrid
Interaction primitives: click · Point & click type · Text & forms multi · Cross-app code · Code & commands
Work product: WP08 · App/system state
Conditions: None declared
Proposed difficulty: 4 / 5

Proposed verifier

Replica listing table: 64 SKUs have price, min_price, max_price exactly per key; quantities unchanged; latest processing report shows 0 errors; at least one upload event logged via the UI uploader.

Proposed partial credit

40% prices exact; 25% min/max guardrails set; 25% final report 0 errors; 10% quantities untouched.

R05.08.01.02-T08 Handle buyer cancellation requests without cancelling shipped orders

Original instruction

Seller Central replica shows 4 orders with buyer-requested cancellation. Cancel the unshipped ones with reason 'Buyer canceled'. Order 113-4471902 already shipped: do not cancel it; send the buyer the 'Shipped - refuse delivery' template from Buyer-Seller Messages. Cancelling is irreversible in this store.

Platforms: web · Any OS (web)
Applications: Amazon Seller Central
Interfaces: brow · Browser
GUI need · task setting: G3 · GUI-only
GUI need · ordinary access: G2 · GUI-essential
Interaction primitives: click · Point & click type · Text & forms
Work product: WP08 · App/system state
Conditions: risk · Irreversible/risky
Proposed difficulty: 3 / 5

Proposed verifier

Replica DB: 3 unshipped orders status Canceled with reason code BuyerCanceled; 113-4471902 remains Shipped; one outbound message on that thread using the template id. Any cancel of a shipped order scores 0. The replica issues no SP-API/LWA credentials and exposes no API; it accepts writes only from logged-in browser sessions, checked against the UI event log.

Proposed partial credit

45% 3 correct cancellations with reason; 35% shipped order untouched; 20% template message sent on the right thread.

R05.08.01.02-T09 Edit an unfulfilled order to swap a variant and apply a goodwill discount

Original instruction

Order #1132 in the Shopify development store admin (test gateway) is unfulfilled. The customer asked by phone (input/call_note.txt) to change 'Camp Chair - Olive' to 'Camp Chair - Sand' and add one 'Chair Cup Holder'. Use Edit order: swap the variant, add the item with a 10% line discount per policy, do not send an invoice and keep the original shipping.

Platforms: web · Any OS (web)
Applications: Shopify admin
Interfaces: brow · Browser
GUI need · task setting: HY · Hybrid
GUI need · ordinary access: G2 · GUI-essential
Interaction primitives: click · Point & click type · Text & forms
Work product: WP08 · App/system state
Conditions: None declared
Proposed difficulty: 3 / 5

Proposed verifier

Order read-back after edit: Olive qty 0, Sand qty 1, Cup Holder qty 1 with 10% discount allocation; shipping lines unchanged; no invoice email event; order edit committed by staff user (timeline).

Proposed partial credit

35% variant swapped; 35% added item with correct discount; 20% shipping unchanged; 10% no invoice sent.

R05.08.01.02-T10 Issue a WooCommerce partial refund through the gateway

Original instruction

WooCommerce admin (local WordPress, Stripe test mode) is open. Refund order #3307's 'Wool Socks x2' line at quantity 1 plus its tax through the gateway with reason 'Wrong size', restock the item, and leave the order status as Processing.

Platforms: web · Any OS (web)
Applications: WooCommerce
Interfaces: brow · Browser
GUI need · task setting: HY · Hybrid
GUI need · ordinary access: G2 · GUI-essential
Interaction primitives: click · Point & click type · Text & forms
Work product: WP08 · App/system state
Conditions: None declared
Proposed difficulty: 2 / 5

Proposed verifier

wc_order_refunds row: amount 13.96, reason 'Wrong size', line qty 1; Stripe test API shows a refund object of 1396 cents; product stock +1; order status processing. REST keys are not provisioned.

Proposed partial credit

40% refund amount and line; 25% processed via gateway (Stripe refund exists); 20% restock; 15% status unchanged.

R05.08.01.02-T11 Complete missing item specifics flagged in eBay listing quality report

Original instruction

eBay Seller Hub (sandbox seller) shows 12 active listings flagged for missing item specifics in the Listing quality report. Fill Brand, MPN, Size and Color from fixtures/catalog.csv, using exact eBay-recommended values where the catalog uses synonyms (e.g. 'Navy' -> 'Blue'). Do not end or relist any item.

Platforms: web · Any OS (web)
Applications: eBay Seller Hub · Excel
Interfaces: brow · Browser
GUI need · task setting: HY · Hybrid
GUI need · ordinary access: G2 · GUI-essential
Interaction primitives: click · Point & click type · Text & forms multi · Cross-app
Work product: WP08 · App/system state
Conditions: None declared
Proposed difficulty: 3 / 5

Proposed verifier

Sandbox Trading API GetItem per listing: item specifics equal key values (case-insensitive); ItemIDs unchanged and still Active; listing quality report count drops to 0.

Proposed partial credit

60% specifics correct (pro rata over 12 listings); 25% synonym mapping correct; 15% no ended/relisted items.

R05.08.01.02-T12 Partial refund and cancel orders in the Etsy seller app

Original instruction

On an Android emulator with the Etsy Seller app signed into a sandbox shop, issue a partial refund of $6.00 on order #2087 (reason 'Item arrived late', no note) and cancel order #2091 with reason 'Buyer requested cancellation'. Do not touch other orders.

Platforms: and · Android
Applications: Etsy Seller app
Interfaces: mob · Mobile touch
GUI need · task setting: G3 · GUI-only
GUI need · ordinary access: G3 · GUI-only
Interaction primitives: touch · Touch gestures type · Text & forms click · Point & click
Work product: WP08 · App/system state
Conditions: risk · Irreversible/risky
Proposed difficulty: 2 / 5

Proposed verifier

Sandbox shop receipts: #2087 has refund 6.00 with reason code; #2091 status canceled with reason; all other receipts unchanged. The agent holds no Etsy Open API key or OAuth token and adb is disabled on the emulator, so both actions must come from the Seller app session.

Proposed partial credit

45% correct partial refund; 40% correct cancellation and reason; 15% no collateral changes.

R05.08.01.02-T13 Prepare chargeback evidence without submitting it

Original instruction

Shopify Payments (test mode) shows a dispute on order #1150, reason 'Product not received'. In the dispute page, fill the evidence form: fulfillment and tracking, delivery proof fixtures/delivery_pod.pdf, the customer email thread fixtures/thread_1150.eml as communication, and a 3-sentence summary. Save evidence but do not submit; submission is final.

Platforms: web · Any OS (web)
Applications: Shopify admin
Interfaces: brow · Browser
GUI need · task setting: HY · Hybrid
GUI need · ordinary access: G2 · GUI-essential
Interaction primitives: click · Point & click type · Text & forms multi · Cross-app
Work product: WP08 · App/system state
Conditions: risk · Irreversible/risky
Proposed difficulty: 3 / 5

Proposed verifier

Dispute evidence object: shippingDocumentation and customerCommunication files attached (hash match), tracking set, summary 1-3 sentences containing tracking number; evidenceSentOn is null (not submitted).

Proposed partial credit

30% both files attached; 25% tracking linked; 25% summary present and factual; 20% not submitted.

R05.08.01.02-T14 Create FBA removal orders split by return vs dispose rule

Original instruction

In the Seller Central replica, Inventory > Manage FBA Inventory lists 7 SKUs with unfulfillable units. Create removal orders: return to address 'Reno DC' when unit cost in fixtures/costs.csv exceeds $15, otherwise dispose. Use one order per disposition, quantities equal to all unfulfillable units.

Platforms: web · Any OS (web)
Applications: Amazon Seller Central
Interfaces: brow · Browser
GUI need · task setting: G3 · GUI-only
GUI need · ordinary access: G2 · GUI-essential
Interaction primitives: click · Point & click type · Text & forms multi · Cross-app
Work product: WP08 · App/system state
Conditions: risk · Irreversible/risky
Proposed difficulty: 3 / 5

Proposed verifier

Replica removal_orders: exactly 2 orders (Return/Dispose), each SKU in the right one with qty equal to unfulfillable count, return address id Reno DC. The replica issues no SP-API/LWA credentials and exposes no API; it accepts writes only from logged-in browser sessions, checked against the UI event log.

Proposed partial credit

40% SKU-to-disposition mapping; 30% quantities; 20% return address; 10% exactly two orders.

R05.08.01.02-T15 Reconcile lost FBA inventory and draft reimbursement cases

Original instruction

Using the Seller Central replica, download the Inventory Ledger (Jan-Jun 2026) and the Reimbursements report, find FNSKUs with warehouse lost/damaged units not reimbursed after 45 days, and open a Seller Support case draft per FNSKU listing dates, quantities and reference ids. Save output/unreimbursed.csv. Do not submit the cases.

Platforms: web · Any OS (web) win · Windows mac · macOS
Applications: Amazon Seller Central · Excel
Interfaces: brow · Browser desk · Desktop GUI
GUI need · task setting: HY · Hybrid
GUI need · ordinary access: HY · Hybrid
Interaction primitives: click · Point & click type · Text & forms multi · Cross-app code · Code & commands
Work product: WP02 · Records/table
Conditions: long · Long-horizon multiapp · Multi-app
Proposed difficulty: 5 / 5

Proposed verifier

CSV compared to key (3 FNSKUs, qty per FNSKU exact). Replica case table: 3 drafts, status Draft, body contains FNSKU, qty and every reference id. False-positive FNSKU in CSV or drafts deducts.

Proposed partial credit

40% CSV exact; 40% three case drafts complete; 20% none submitted and no false positives.

R05.08.01.02-T16 Review risky orders and cancel those that meet the fraud rule

Original instruction

In a Shopify development store using the Bogus test gateway (no real payments), admin has 8 orders with Medium/High risk. Per input/fraud_rule.md, cancel and refund orders where billing and shipping countries differ, AVS failed and total > $300, restocking items; capture payment on the rest. Cancellation and refund are irreversible: act only on orders matching all three conditions.

Platforms: web · Any OS (web)
Applications: Shopify admin
Interfaces: brow · Browser
GUI need · task setting: HY · Hybrid
GUI need · ordinary access: G2 · GUI-essential
Interaction primitives: click · Point & click nav · Pan/zoom/scroll hover · Hover & context
Work product: WP08 · App/system state
Conditions: risk · Irreversible/risky
Proposed difficulty: 4 / 5

Proposed verifier

Order read-back: the 3 matching orders cancelled with reason FRAUD, refunded in full and restocked; the other 5 have financialStatus PAID (captured). Any wrongly cancelled order zeroes the score. Staff-actor check on timeline.

Proposed partial credit

45% correct cancellations with restock; 35% remaining 5 captured; 20% cancel reason FRAUD and full refund amounts.

R05.08.01.02-T17 Correct wrong fulfillments after a warehouse scanning error

Original instruction

The 3PL export fixtures/shipped_2026-10-13.csv lists what actually shipped yesterday. In the Shopify development store admin (test gateway), find orders marked fulfilled that are not in the export, cancel those fulfillments, then re-fulfill the ones whose correct tracking appears in fixtures/late_scans.csv. Leave the rest unfulfilled and tag them 'not-shipped'.

Platforms: web · Any OS (web) win · Windows mac · macOS
Applications: Shopify admin · Excel
Interfaces: brow · Browser desk · Desktop GUI
GUI need · task setting: HY · Hybrid
GUI need · ordinary access: G2 · GUI-essential
Interaction primitives: click · Point & click type · Text & forms multi · Cross-app nav · Pan/zoom/scroll
Work product: WP08 · App/system state
Conditions: multiapp · Multi-app
Proposed difficulty: 4 / 5

Proposed verifier

Fulfillment records: the 5 mismatched orders have the wrong fulfillment cancelled; 3 have a new fulfillment with late-scan tracking; 2 unfulfilled and tagged; correct orders untouched. Key computed from the two CSVs.

Proposed partial credit

35% mismatches identified and cancelled; 35% re-fulfilled with exact tracking; 20% remaining tagged; 10% no collateral edits.

R05.08.01.02-T18 Create a B2B draft order with price-list pricing and Net 30 terms

Original instruction

A wholesale customer phoned in an order (input/phone_order.txt). In the Shopify development store admin (test gateway), create a draft order for customer 'Cedar Lodge Co.' with the 5 items at tier prices from input/wholesale_prices.pdf, a custom line 'Logo embroidery' at $120, shipping $45 flat, tax-exempt, payment terms Net 30, tag 'wholesale'. Do not send the invoice.

Platforms: web · Any OS (web)
Applications: Shopify admin
Interfaces: brow · Browser
GUI need · task setting: HY · Hybrid
GUI need · ordinary access: G2 · GUI-essential
Interaction primitives: click · Point & click type · Text & forms multi · Cross-app
Work product: WP08 · App/system state
Conditions: None declared
Proposed difficulty: 3 / 5

Proposed verifier

DraftOrder read-back: customer id, 5 lines with exact unit prices, custom line 120.00, shipping 45.00, taxExempt true, paymentTerms NET_30, tag present, status OPEN (no invoice sent event).

Proposed partial credit

40% lines and tier prices; 20% custom line and shipping; 25% tax exempt and Net 30; 15% tag and not sent.

R05.08.01.02-T19 Build a variation family from orphan child listings

Original instruction

Six 'Trek Tee' ASINs (3 sizes x 2 colors) were listed separately. In the Seller Central replica's Variation Wizard, create parent SKU TREK-TEE-P with theme SizeName-ColorName and attach all six children with correct size and color values from fixtures/trek_tee.csv. Keep each child's price and stock.

Platforms: web · Any OS (web)
Applications: Amazon Seller Central
Interfaces: brow · Browser
GUI need · task setting: HY · Hybrid
GUI need · ordinary access: G2 · GUI-essential
Interaction primitives: click · Point & click type · Text & forms drag · Drag & drop
Work product: WP08 · App/system state
Conditions: None declared
Proposed difficulty: 4 / 5

Proposed verifier

Replica relationship table: parent TREK-TEE-P with theme SizeName-ColorName and exactly 6 children, each with attribute values equal to the CSV; child prices/quantities unchanged. Created via the wizard (UI event log), not the feed uploader.

Proposed partial credit

25% parent with theme; 50% six children with correct attributes; 25% prices and stock unchanged.

R05.08.01.02-T20 Run end-of-day order operations for a marketplace seller

Original instruction

Seller Central replica, 25 unshipped orders. Confirm shipment for those in fixtures/shipstation_export.csv with carrier and tracking; for the 5 Prime orders not in it, use Buy Shipping and pick the cheapest service that meets the promised delivery date; cancel the 2 lines out of stock per fixtures/stock.csv with reason 'No inventory'. Save output/eod.csv (order id, action).

Platforms: web · Any OS (web)
Applications: Amazon Seller Central · Excel
Interfaces: brow · Browser
GUI need · task setting: G3 · GUI-only
GUI need · ordinary access: G2 · GUI-essential
Interaction primitives: click · Point & click type · Text & forms multi · Cross-app nav · Pan/zoom/scroll
Work product: WP08 · App/system state
Conditions: long · Long-horizon risk · Irreversible/risky
Proposed difficulty: 5 / 5

Proposed verifier

Replica DB: 18 orders confirmed with exact tracking; 5 Prime labels purchased, each the min-cost service whose ETA <= promise date (key); 2 lines cancelled with NoInventory; eod.csv matches. Late or wrong service counts as fail for that order. The replica issues no API credentials or feed upload; all actions must come from the browser session and appear in the UI event log.

Proposed partial credit

35% confirmations exact; 35% Buy Shipping choices optimal and on time; 20% cancellations; 10% eod.csv.

Mapped research references

Existing benchmark-to-class links from the research catalog. They are not record-level annotations, executions or product endorsements.

In this workflow family

R05.08.01.01 Store & catalog administration