--- name: price-drop-watch description: Compare the current price of a specific product across major retailers (Amazon, Walmart, eBay, Target, Home Depot) plus Google Shopping, side-by-side, and flag the better deal. Use when the user asks "where is X cheapest", "Amazon vs Walmart price for X", "find the lowest price on X", or wants a quick price diff between retailers before buying. Returns a compact comparison table, not a full review. license: MIT metadata: author: scavio version: "1.1" requires: scavio-mcp homepage: https://github.com/scavio-ai/hermes-agent --- # Price drop watch Fast price comparison for a single product across the retailers Scavio covers. No reviews, no sentiment, just price plus availability plus a clear verdict. If the user wants reviews and sentiment in addition to price, route to `product-deep-research` instead. For travel pricing (flights, hotels, stays) this skill is the wrong shape - Scavio's Google Flights, Google Hotels, Booking.com and Airbnb endpoints have their own date and occupancy inputs. ## When to use - "Is X cheaper on Amazon or Walmart?" - "Find the lowest price on X." - "Quick price check for X." - "Where should I buy X?" ## Inputs - **Product**: name or SKU. Be specific. "iPhone 16 Pro 256GB Natural Titanium" beats "iPhone". - **Retailers** (optional): defaults to Amazon and Walmart. Widen when the category warrants it - see the table under Procedure step 1. ## Retailer coverage Pick sources by category rather than querying everything. Each extra source costs credits and adds a normalization step. | Source | Good for | Credits per request | | --- | --- | --- | | Amazon | general, default | 1 | | Walmart | general, groceries, default | 1 | | eBay | refurbished, discontinued, open-box | 1 | | Target | household, apparel, toys | 1 | | Home Depot | tools, appliances, building materials | 2 | | Google Shopping | merchants none of the above carry | 1 | If the product sells mainly on a merchant Scavio has no dedicated endpoint for, fall back to Google Shopping to find the listing, then use the Extract endpoint on that product URL to pull the price off the page. Extract is tier-priced rather than flat - check the docs at https://scavio.dev/docs before budgeting a run around it. ## Procedure 1. **Choose sources.** Default to Amazon and Walmart. Add from the table above when the category clearly lives elsewhere (power tools, refurbs, niche merchants). Two sources is usually enough for a verdict. 2. **Search each source** via Scavio for the same product string. Capture: title, price, availability, rating and review count where the source returns them, seller (first-party vs third-party), shipping, URL. Take the top first-party listing where available. 3. **Normalize.** Verify the listings are the same product configuration (same model, color, capacity, bundle vs bare unit, new vs refurbished). If they differ, flag the mismatch and pick the closest comparable on each side. 4. **Compute the diff** against the cheapest in-stock listing and output the table below. ## Output format ```markdown # price check | Retailer | Price | Stock | Seller | Shipping | | --- | --- | --- | --- | --- | | | | | | | | | | | | | **Cheaper at by (%).** - : - : ``` Every cell comes from the API response. Do not fill a gap with a remembered or estimated price - if a source returned no price, write "not returned" and say so in the caveat. ## Edge cases - **Product not listed on one of the sources.** Show the sources that have it and mark the others "not listed". Consider adding Google Shopping before concluding the product is unavailable. - **Third-party-only Amazon or eBay listing.** Flag it; third-party prices fluctuate and may include markups. - **Condition mismatch.** eBay results are often refurbished or open-box. Never compare a refurbished price against a new one without labelling both. - **Bundle mismatch** (e.g., one listing includes a case, the other doesn't). Normalize to bare unit if possible, otherwise flag. - **Stock outage.** "In stock at " beats "out of stock at a lower price". Recommend the available one even if nominally pricier. - **Tool not registered.** The hosted MCP server at https://mcp.scavio.dev/mcp exposes a default subset of its 191 tools. If a retailer's tool is missing from the session, widen the surface with the `x-scavio-platforms` header (or the `SCAVIO_PLATFORMS` env var locally). It is additive - `default,` keeps the default set and adds the named one. Check https://scavio.dev/docs for the platform identifiers; `all` registers all 191 tools but ships a large payload into every session. ## Cost Credit cost is per request and **varies by platform** - it is not a flat rate across Scavio. The per-request costs for the sources this skill uses are in the Retailer coverage table above. A typical two-source run (Amazon plus Walmart) is 2 credits, plus 1 per extra search if the first pass needs a retry or a follow-up lookup. Adding Home Depot costs 2 rather than 1. New accounts get 50 credits on signup, one-time, no credit card required - enough for roughly 25 default two-source runs. Sign up at https://dashboard.scavio.dev/sign-up.