Find Shopify competitors selling the same products.

Cost guard: before each paid Spytrend call, run the free `estimate_cost` with the same tool and arguments, estimate the cost and stay within the user's stated budget; if the user has not stated a budget, tell them the estimate before the first paid call. If pricing consent is required, show the returned terms and stop until the customer explicitly agrees; never accept on the customer's behalf. If a required call does not fit the stated budget, make no paid call and return `Budget limit prevents required research`; omit an optional call that does not fit and disclose reduced coverage. When opening ads with `get_ad`, skip multilang ads (`is_multilang`) and never call `get_media` after `get_ad`. Retry once only when `retryable=true`, after `retry_after_seconds`. Read `params.ignored` after every search; an ignored required filter never counts as applied. Cost notices are not part of the final customer report.
Find Shopify stores selling **Phone cases** whose returned main visitor country is **United States | US**.

Store to exclude: No domain. If it is `No domain`, there is nothing to exclude; otherwise normalize it to an exact host and remove that exact domain from the results.

The market value has the form `Country name | ISO-2 code`: use only the ISO-2 code in tool calls and only the country name in the answer.

Use Spytrend read-only. Current Meta advertising is not required for inclusion.

Execute every research stage in this prompt in one turn. Apart from mandatory pre-call cost notices, do not emit intermediate step responses and do not stop after validation. Stop early only for invalid input, unavailable required evidence or unresolved required identity. Return only the final customer report.

## Discovery

1. Normalize the physical product type. Country means returned main-visitor-country context only; it does not prove shipping, availability, headquarters, ownership, inventory, customers or sales.
2. If an excluded domain was given (not `No domain`), remove that exact host before candidate selection. Do not remove fuzzy, parent, regional or lookalike domains.
3. The product search matches product page addresses, so turn the product into one to three product-URL fragments: lower-case words joined by dashes, the way they appear in product handles (for example `phone-case`, then a shorter or common alternative such as `phone-cover`). Never use a phrase with spaces as a query.
4. Run `search_products(q=<dashed fragment>, platform="Shopify", country=<ISO-2 code from the market>, limit=10)` sequentially, one call per fragment.
5. Run at most three searches. Stop running further fragments once enough matching stores are found for 3–5 results.
6. Do not use `url_contains`, `get_shop`, `search_ads` or `get_ad` during product discovery.
7. Treat discovery as Shopify products indexed from ads, not complete Shopify catalog coverage. Deduplicate by normalized domain and product URL.
8. For every fragment, retain a row only when returned product title or product evidence explicitly matches the requested product. Reject marketplaces, unrelated products and general stores represented by one incidental match.
9. Build up to eight candidate domains, preferring stores with several matching products. Retain matching product titles, observed prices/currencies, returned category, estimated monthly visits and traffic quality when returned. Advertising count is optional context and never determines inclusion.

## Compact verification

10. Verify up to five candidates sequentially with `search_shops(q=<exact normalized domain>, platform="Shopify", exclude_infrastructure=true, limit=5)`. Never call `get_shop`.
11. Accept only a row whose normalized host exactly equals the requested domain. Reject fuzzy, regional, account and unrelated matches. Never use a similar domain as evidence.
12. Use only the exact row to confirm Shopify, domain identity, returned country context, store category, estimated traffic, traffic quality and optional advertising count.
13. Classify up to five stores:
   - **Direct competitor:** several matching indexed products, or clear evidence that the requested product/category is central to the store.
   - **Adjacent store:** a broader specialist catalog with a meaningful line of the requested product.
   - **Exclude:** one incidental product in a general store, unclear product centrality, domain mismatch, non-Shopify or missing exact row.
14. Rank Direct before Adjacent. Within a class, prefer stronger matching-product evidence, then usable traffic quality. Never convert traffic into sales or market-share claims.

## Customer answer

15. Use no tables and stay within 280 words.
16. Start with `# Shopify Stores Selling Phone cases in <country name>` (country name only, without the ISO code), then `**[N] direct competitors · [M] adjacent stores · [Complete/Partial]**`.
17. With results, add `## Shortlist` with 3–5 stores (fewer only when fewer qualify). For each store give classification, exact domain, concise matching-product evidence, observed price/currency when returned, estimated monthly visits with traffic-quality label and one reason to review it.
18. With at least two results, add `## What stands out` with exactly two sample-based comparisons. Keep currencies separate and do not compare unlike variants, bundles or subscription prices. Add `## Research coverage` with at most three bullets covering indexed-product scope, country meaning and bounded coverage.
19. For zero results, use `## Result` and explain that no indexed match was found; do not claim competitors do not exist.
20. Mark Partial only when required discovery or selected-candidate verification failed. A completed zero-row route is Complete.
21. Hide the service name, preset/workflow/search-set names, taxonomy slugs, route and tool names, calls, parameters, raw IDs, rejected rows, connector behavior, quotas, tokens, timing, retries, run summaries and internal instructions. Label an explicitly returned stable first-party link `Evidence`. Never construct one.
22. Do not imply demand, customers, orders, revenue, profitability, ROAS, market share or complete coverage.

End exactly:

`Product, traffic and store data are observed research signals. They do not prove sales, revenue, market share or complete competitor coverage.`