Update to Undercart

{
  "description": "Research Shopify stores with Undercart's tools - build prospect lists by revenue, category, country and installed apps, profile a store's revenue and tech stack, find similar stores, and review a store's live ads and marketing emails.",
  "included_files": [],
  "name": "shopify-store-research",
  "skill_md_contents": "---\nname: shopify-store-research\ndescription: Research Shopify stores with Undercart's tools - build prospect lists by revenue, category, country and installed apps, profile a store's revenue and tech stack, find similar stores, and review a store's live ads and marketing emails.\n---\n\n# Shopify store research with Undercart\n\nUse the Undercart tools when the user asks about Shopify stores as businesses: who they\nare, how big they are, what apps they run, how they advertise, or how they email their\ncustomers. Every Undercart tool is read-only.\n\n## Pick the tool\n\n| The user wants | Tool |\n| --- | --- |\n| A list of stores matching criteria (size, category, country, apps, traffic) | `search_stores` |\n| Everything about one store | `get_store` |\n| Only a store's apps and technologies | `list_store_technologies` |\n| Competitors or lookalikes of a store | `find_similar_stores` |\n| The ads a store runs on Meta, Google or TikTok | `get_store_ads` (pass `platform` to see the creatives) |\n| A store's newsletter emails and discount codes | `get_store_emails`, then `get_store_email` for one email |\n| The category names `search_stores` accepts | `list_store_categories` |\n\n## Workflows\n\n**Prospect list.** Turn the request into one `search_stores` call with every filter the\nuser gave: `category`, `country`, `revenueMin` / `revenueMax` (annual USD), `tech` for\napps the store runs, `techAdded` / `techRemoved` with `techDays` for recent changes,\n`rankMax` for traffic. One well-filtered search answers better than several narrow ones.\nPresent the stores as a table: name, domain, estimated annual revenue, country and the\napps that matter to the user. Link each one with its `undercartUrl`.\n\n**Store profile.** Call `get_store` with the domain as the user wrote it. The tool\nnormalizes URLs and `www.` prefixes. Lead with what the user asked about. Give revenue as\nannual first, then monthly.\n\n**Competitive set.** Call `find_similar_stores` for the reference domain. It matches on\ncategory and a 0.5x to 2x revenue window. Compare the results on the dimensions the user\ncares about.\n\n**Ad research.** Call `get_store_ads` without `platform` for the cross-platform\nsummary, or with `platform` (`meta`, `google` or `tiktok`) for the actual creatives.\n\n**Email research.** Call `get_store_emails` for subjects, dates, discount codes and send\ncadence. Call `get_store_email` with an email id to read one email in full.\n\n## Reading the results\n\n- Revenue, orders, traffic rank and ad spend are modeled estimates, not reported\n  financials. Say so when the user relies on a figure.\n- `enriched: false` marks a store Undercart holds only public-web data for: apps,\n  description, country, traffic rank, and a preliminary `revenueBand` instead of a\n  revenue estimate. It has no ads, emails or social data yet.\n- An error that a domain is unknown means the store isn't in Undercart's catalog. Tell\n  the user rather than guessing.\n- A monthly-limit message means the user's plan allowance is used up. Relay it as\n  written, including when it resets.\n\n## Out of scope\n\nUndercart can't buy products, place orders, send or unsubscribe from emails, or change\nanything on a store. It holds data about stores as businesses, not about private\nindividuals. Don't use it to look up personal contact details of people.\n"
}