agentsclimarketplace

Ucp product discovery

Skill fxp/agentic-commerce-skills/skills/01-discovery/ucp-product-discovery

Make a merchant's products discoverable to commerce agents under UCP — by publishing an accurate Merchant Center product feed and a /.well-known/ucp profile so Google AI Mode / Gemini can surface buyable items. Use when setting up the discovery side of UCP, or when reasoning about how an agent finds products before a checkout session. NOTE: core UCP has NO agent-facing catalog "search" REST endpoint — discovery is feed- and index-based.From its SKILL.md

Install
npx -y skills add fxp/agentic-commerce-skills --skill ucp-product-discovery

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its file declares

Copied from the file, not written here

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

4.7 KB, ~1.1k tokens by cl100k_base, as published. Nobody here has run it

UCP Product Discovery

中文摘要:UCP 的"发现"环节靠商家发布 Merchant Center 商品 feed + /.well-known/ucp profile,由 Google(AI Mode/Gemini)索引后把可购买商品呈现给 agent。核心 UCP 没有给 agent 调用的"目录搜索"REST 端点——发现是 feed/索引驱动的。

⚠️ Accuracy note / 准确性说明

An earlier draft of this repo modeled discovery as an agent calling a merchant catalog.search REST endpoint. That is not how core UCP works. The UCP technical spec defines capabilities for checkout (incl. cart management), fulfillment, discount, order, and common identity_linkingno standalone catalog/search/cart spec. Product discovery happens through the merchant's product feed (Merchant Center) which Google indexes for AI surfaces. 早期草稿把发现建模成 agent 调商家搜索接口,这是错的;实际靠商品 feed。

When to use / 何时使用

  • A merchant wants their products eligible to appear in agentic shopping on Google.
  • You're reasoning about the step before a checkout session: how the agent got item.ids.
  • You're auditing whether a merchant's feed + UCP profile are discovery-ready.

UCP mapping / UCP 映射

Discovery is enabled by two published artifacts, not a callable capability:

  1. Product feed — submitted via Merchant Center (title, price, availability, GTIN, etc.). Google indexes it; agents on AI Mode / Gemini surface matching items to users.
  2. UCP profile at /.well-known/ucp (public, no auth) — declares the merchant's ucp.services["dev.ucp.shopping"] endpoint + ucp.capabilities, so once a user picks a product, the agent knows where to open a checkout session.

The product id a user lands on becomes line_items[].item.id in the checkout session.

Inputs & Outputs / 输入与输出

Merchant-side inputs: a complete product feed + a valid /.well-known/ucp profile.

What the agent actually receives (from Google's index, not a merchant API):

{ "item": { "id": "sku_8821", "title": "Trail Runner GTX", "price": 11900 },
  "merchant_profile": "https://merchant.example/.well-known/ucp" }

price is an integer in the currency's minor units (e.g. cents). 价格为最小货币单位整数。

Workflow / 工作流 (merchant enablement)

  1. Publish the product feed in Merchant Center; keep availability and price fresh — stale data fails at checkout, not discovery.
  2. Serve /.well-known/ucp declaring ucp.version, ucp.services (transport rest, endpoint, schema), and ucp.capabilities (e.g. dev.ucp.shopping.checkout).
  3. Verify eligibility in Merchant Center; fix feed disapprovals.
  4. Hand off: the discovered item.id seeds ucp-checkout-session.

Edge cases & failure modes / 边界与失败

  • Treating discovery as a live API → it isn't; don't design an agent to "search the merchant". Search/ranking is Google's; the merchant supplies data.
  • Feed/profile mismatch → product indexed but profile lacks checkout capability → not buyable.
  • Stale price/availability → discovery may show it; checkout re-validates and may reject.
  • Region/currency → feed targeting controls where items are eligible.

Worked example / 示例

Merchant onboards "Trail Runner GTX".

  1. Feed row: id=sku_8821, price=11900 USD, availability=in_stock, gtin=….
  2. /.well-known/ucp declares dev.ucp.shopping REST endpoint https://merchant.example/ucp/v1.
  3. User on Gemini asks for trail shoes → Google surfaces sku_8821 → agent opens a checkout session at the declared endpoint with line_items:[{item:{id:"sku_8821"}, quantity:1}].

References / 参考

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Gives 0 of the 12 instructions most product growth skills give in ~1.1k tokens

Counted across 728 of the 1,010 authors here whose files we hold, read 2026-08-07

  • Read product marketing context before asking questionsin 24 of 728, across 18 files
  • Define the ideal customer profilein 21 of 728, across 3 files
  • Document a rollback plan before deploymentin 21 of 728, across 12 files
  • Analyze the codebase to understand the productin 19 of 728, across 1 file
  • Ask clarifying questions about the value propositionin 19 of 728, across 1 file
  • Search for companies matching the criteriain 19 of 728, across 1 file
  • Look for signals of immediate needin 19 of 728, across 1 file
  • Assign a fit score from one to tenin 19 of 728, across 1 file
  • Identify the target decision-maker rolein 19 of 728, across 1 file
  • Suggest a personalized contact strategyin 19 of 728, across 1 file
  • Provide conversation starters for outreachin 19 of 728, across 1 file
  • Format results in a scannable markdown templatein 19 of 728, across 1 file

Said here and by no other author read

  • Publish the product feed in Merchant Center
  • Keep availability and price data fresh
  • Serve a valid .well-known/ucp profile
  • Declare ucp version, services, and capabilities in the profile
  • Verify eligibility and fix feed disapprovals
  • Use discovered item ids to seed checkout sessions

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 326,367. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.