Check the input, then compare the complete deliverables
Explore the original materials, item-by-item outcomes and full report below. Attempt the input independently before using the reference to find omissions or disagreements.
Listings, locations and prices are fictional. No real platform search, user geolocation, seller contact, offer or payment occurred.
Original synthetic material. Try the input independently, then compare the reference results.
Original synthetic input
{
"synthetic": true,
"case": "marketplace-watch",
"data": {
"criteria": {
"brands": [
"A",
"B"
],
"max_price_usd": 300,
"max_distance_miles": 30,
"condition": "Seller claims good or better"
},
"snapshots": [
{
"row": "L01",
"listing_id": "item1",
"brand": "A",
"price": 280,
"distance": 12,
"condition_claim": "good",
"state": "available",
"observed": "2026-10-16T09:00:00+08:00"
},
{
"row": "L02",
"listing_id": "item2",
"brand": "B",
"price": 320,
"distance": 10,
"condition_claim": "good",
"state": "available"
},
{
"row": "L03",
"listing_id": "item3",
"brand": "A",
"price": 250,
"distance": null,
"condition_claim": "good",
"state": "available"
},
{
"row": "L04",
"listing_id": "item1",
"brand": "A",
"price": 270,
"distance": 12,
"condition_claim": "good",
"state": "available",
"observed": "2026-10-16T18:00:00+08:00"
},
{
"row": "L05",
"listing_id": "item4",
"brand": "B",
"price": 200,
"distance": 8,
"condition_claim": "good",
"state": "sold"
}
]
}
}Processing rules
Merge snapshots by listing_id and retain change history. A price drop is an update, not a new listing. Unknown distance is pending, not 0. Exclude sold/over-budget items. Condition is a seller claim, not an inspection. Deliver candidates/questions without contacting or transacting.
Complete reference results under this tutorial’s rules; nothing was written to a third-party service.Scroll horizontally within the table if needed.
| Listing | Latest facts | Filter result | Needs verification |
|---|---|---|---|
| item1 | USD 280→270; 12mi; available | One candidate; price update | good is a seller claim; inspect item/accessories |
| item2 | USD 320; 10mi | Over budget; excluded | Do not raise budget automatically |
| item3 | USD 250; distance unknown | Pending | Do not assume0mi |
| item4 | USD 200; 8mi; sold | Excluded | Low price is not availability |
item1 drops byUSD10 from 280 to 270. Keep one identity and one price-change event.
Accessories, demonstration, actual condition and repair history remain unknown.
Marketplace report — synthetic snapshots 5 rows, 4 unique listings. item 1 is the only complete match: latestUSD270, 12mi, price drop 10; condition is only a seller claim. item 2 over budget; item3 distance pending; item4 sold. No seller contact, offers, geolocation or payment. Matching price does not verify quality or transaction safety.
Prepare the materials and check your tools
Start with the practice files
- Download five snapshots and criteria.
- L01/L04 refer to item1; deduplicate before counting.
- Use USD and miles consistently.
Before using real services
- Let the user choose city and radius; infer no address.
- Confirm platform/account scope and user takeover for verification.
- Record necessary public listing fields only.
Complete one file-based exercise first. For real services, verify the actual account, available tools and operation scope, then follow the acceptance check at each stage.
Download the original input and rules ↓Execution and takeover boundaries
The actions below require the agreed inputs, accounts and scope. Configuring a Routine does not expand that authorization.
Can run automatically
Filter agreed snapshots by criteria; deduplicate by listing_id and separate new items, price drops, removals and missing data.
Requires confirmation
Confirm the live platform, city, radius and Routine notification scope. Do not infer user location; seller contact, bids and payment need separate approval.
Take over after a failure
Let the user handle login verification; read failure is not no change. Pause monitoring, preserve coverage and resume deduplication by listing identity after manual checks.
Fix budget, units and required conditions
Do not widen distance/model requirements because a price looks attractive. Location must come from the user’s chosen scope.
Restate brandsA/B, USD 300 maximum, 30mi radius and seller condition requirement. List fields and missing-value rules. Do not search, obtain user location or contact sellers.Explicit selection criteria.
Units fixed and budget not widened.
Merge repeated observations by identity
Titles can change. A stable listing id lets you report a discount without counting a second new candidate.
Group5 snapshots by listing_id, retaining original rows/times. L01/L04 form item1’s280→270 history. Keep both values and do not create two listings from changed titles/prices.Four identities with history.
item1 counts once and theUSD10 change is traceable.
Separate unknown fields from nonmatches
Insufficient information is neither a match nor a failure. Sold items leave candidates even when cheap.
Check price, distance, brand, state and condition claim. Keep item3 distance pending; exclude item2 over budget and item4 sold. Infer no inspected condition.Four filter outcomes.
Only item1 completely matches; item3 is not treated as0mi.
Explain candidate reasons and inspection questions
A report helps decide what to verify. It cannot establish seller trust or item quality from a snapshot.
Write summary, candidate changes, exclusions and questions. Label good as seller-stated; accessories/faults/repairs remain unknown. Keep a local report; do not contact, offer or pay.A report and verification questions.
Price fit is not purchase endorsement or quality verification.
Observe a bounded real scope with timestamps
Platform ranking and login state affect visible results. A small read is not a whole-city cheapest-price survey.
Read at most 10 listings from <specified platform/city/criteria>. Record listing_id, time, visible fields and failures. Do not guess distance. Let the user handle verification. Do not contact/transact or collect unrelated personal information.A bounded real report or blocker.
Coverage explicit; no lowest-everywhere or inspected-quality claim.
Distinguish discoveries, discounts and removal
Persist identity state so repeat sightings are updates. A failed read is not no new listings.
Draft a listing-watch Skill with criteria, listing_id, time, deduplication, price changes, sold states and gaps. Propose a Routine only. Notify the current user on new matches/discounts/status changes; never message sellers, bid or pay.Rerun/change rules.
Repeated sightings are not new; sold listings do not stay available.
Find the evidence before choosing the next action
The same item appears new daily
Use stable identity and separate title/price updates.
Distance is unreadable
Mark pending without assuming nearby.
The seller says good
Retain it as a claim and list inspection questions.
No new items or failed read?
Check run status and coverage; report those states separately.
Check every item before finishing
These page-local checkboxes reset on refresh. Checking them does not verify a real service connection.
Checked 0 / 6 items
Download the inputs separately from the reference answers
Start with input and rules without showing the reference answers. After acceptance, save the report and complete prompt set.
source-data.json↓Rules and scopeDefinitions, exceptions and delivery requirements.workflow-rules.md↓Item-by-item reference resultsCompare every input with its outcome.expected-results.csv↓Complete reference reportResults, evidence and missing information.reference-report.md↓All stage prompts6 instructions with expected outputs and acceptance checks.practice-prompts.txt↓Check documented capabilities and exercise results separately
Content updated: · Editing a page does not reverify its sources or live workflow.
Documentation review: Unverified · no review date recorded
No case-specific documentation review is recorded. References below are provided for review and do not establish verification.
Live workflow verification: Unverified
Live accounts, external writes, scheduled execution and long-term operation are unverified. Synthetic exercises and page checkboxes do not establish live verification.
Listings, locations and prices are fictional. No real platform search, user geolocation, seller contact, offer or payment occurred.
- Grok Bot: apps and the shared computer ↗
Check connected accounts, available tools, visibility and user takeover.
- Grok Bot: apps and the shared computer ↗
Check connected accounts, available tools, visibility and user takeover.
- Grok Bot: Skills and Routines ↗
Validate a manual run before saving a method or scheduling it.
Try another small batch of your own material
Retain original input and operation records. Verify that each result can be traced before adding connections or recurring runs.