# Vendor facets — Shopify. A commerce platform that GENERATES agent surfaces on every store:
# UCP MCP at /api/ucp/mcp (successor to Storefront MCP at /api/mcp), a /.well-known/ucp profile,
# an auto-generated /llms.txt mirroring /agents.md, and Catalog syndication to AI channels. They
# are one shape across millions of stores, so the rubric grades them `platform` and `derived`,
# not as the merchant's own work. 0.23.0 reverses #188/#123's crediting of these as the brand's own.
vendor: shopify
name: Shopify
website: https://www.shopify.com
areas:
- hosting
registry_keys:
- shopify
rubric_schema_version: 0.22.0
generated: '2026-09-25'
features_refreshed: '2026-09-25'
basis: capability
summary: >-
  Every Shopify store gets an MCP endpoint, a UCP commerce profile and an llms.txt without doing anything
  — which is exactly why they earn little: the tools, profile and file are Shopify's, identical across
  stores, so mcp_server and agentic_commerce grade `platform` (0.25) and a generated llms.txt counts as
  derived (0.25). A merchant who hand-writes llms.txt.liquid earns that check as its own work. Catalog
  syndication to ChatGPT, Gemini and Copilot is real distribution that no Kin Score check reads.
features:
- id: ucp-mcp
  name: UCP MCP per store (successor to Storefront MCP)
  description: >-
    Every store answers MCP at https://{shop}/api/ucp/mcp with Shopify-defined catalog, cart and checkout
    tools; the older Storefront MCP at /api/mcp is deprecated for catalog and cart.
  source: https://shopify.dev/docs/apps/build/storefront-mcp/servers/storefront
  tier: all
- id: ucp-profile
  name: UCP business profile at /.well-known/ucp
  description: >-
    The store's UCP profile (version, services, capabilities, payment handlers, signing keys) served at
    {shop}.myshopify.com/.well-known/ucp, published by the platform operating the commerce API.
  source: https://shopify.dev/docs/agents/profiles
  tier: all
- id: llms-txt-auto
  name: Auto-generated /llms.txt and /agents.md
  description: >-
    Shopify serves /llms.txt by mirroring its generated /agents.md; a merchant can override it with an
    llms.txt.liquid or agents.md.liquid theme template it writes.
  source: https://shopify.dev/docs/storefronts/themes/architecture/templates/llms-txt-liquid
  tier: all
- id: catalog
  name: Shopify Catalog for agentic storefronts
  description: >-
    Eligible products are syndicated automatically to AI channels (ChatGPT, Microsoft Copilot, Google
    AI Mode / Gemini, Meta) with per-channel opt-out.
  source: https://help.shopify.com/en/manual/online-sales-channels/agentic-storefronts/products
  tier: all
- id: ucp-platform
  name: UCP (Universal Commerce Protocol) and Global Catalog
  description: >-
    Shopify's agent-commerce stack — UCP negotiation, a Global Catalog across all merchants, checkout
    for trusted agents.
  source: https://shopify.dev/docs/agents
  tier: all
maps:
- feature: ucp-mcp
  check: mcp_server
  layer: agent_readiness
  grade: platform
  note: >-
    One Shopify-defined tool set on every store; 0.22.0 grades that `platform` (0.25), and 0.23.0's platform-generated
    class (#676/#680) marks it explicitly. A merchant earns more only by operating its own server.
  points: 12
  baseline_pass_rate: 0.22
- feature: ucp-profile
  check: agentic_commerce
  layer: agent_readiness
  grade: platform
  note: >-
    The profile is on the store's domain, so a literal read of the grade rule says `self`; its services
    and endpoints are Shopify's, identical across stores, so `platform` (0.25) is the honest grade and
    the one 0.23.0 applies.
  points: 5
  baseline_pass_rate: 0.002
- feature: llms-txt-auto
  check: llms_txt_published
  layer: composite
  credit: 0.25
  note: >-
    The default file is Shopify-generated, not written by the merchant — derived (0.25). A hand-written
    llms.txt.liquid / agents.md.liquid is the merchant's own at 1.0, and that credit is the merchant's
    work, not Shopify's.
  provider_must: Nothing for the derived default; write the Liquid template to earn full credit.
  catalog_pass_rate: 0.351
  facet: discoverability
  points: 4
  baseline_pass_rate: 0.65
earns_nothing:
- feature: catalog
  why: Syndicating products into AI channels is distribution; no check reads where a catalog is listed.
- feature: ucp-profile
  check: well_known_published
  why: >-
    The check wants an api-catalog, security.txt, protected-resource or AAuth document; /.well-known/ucp
    is none of them (it is read by agentic_commerce instead).
out_of_reach:
  checks:
  - well_known_published
  - protected_resource_metadata
  - contract_present
  - sdk_count_1
  - agent_card
  note: >-
    We found no documented way for a store to serve its own api-catalog, security.txt or protected-resource
    document; Storefront GraphQL and UCP are Shopify's contracts, not the merchant's.
unscored_practice:
- feature: llms-txt-auto
  why: >-
    /agents.md as the canonical agent-discovery file (llms.txt falls back to it) is deployed on every
    store and no check reads it.
surface:
  discoverability:
    reachable: 1.0
    total: 54
  agent_readiness:
    reachable: 4.2
    total: 139
hard_rule: >-
  A model, not a score. Adopting this vendor changes a provider's Kin Score only when the provider publishes
  the resulting artifacts on its own surface; nothing here writes a score, and no sponsorship or partnership
  can.
method: searched
source:
- https://help.shopify.com/en/manual/online-sales-channels/agentic-storefronts/products
- https://shopify.dev/docs/agents
- https://shopify.dev/docs/agents/profiles
- https://shopify.dev/docs/apps/build/storefront-mcp/servers/storefront
- https://shopify.dev/docs/storefronts/themes/architecture/templates/llms-txt-liquid
measured:
  cohort:
    method: vendors-catalog.json detections (CNAME / header / URL shape / markup), never a name match
    detected: 17
    in_baseline: 2
  control:
    basis: providers earning contract_present + documentation_present + api_reference_present, minus the
      cohort
    n: 5214
  metric: >-
    cohort_pct / control_pct = mean share of the check's points earned (derived and platform credit weighted),
    x100
  measured_on: '2026-09-25'
  status: 'not measurable: 2 detected customers clear the baseline (need 20)'
simulation:
  simulated_on: '2026-09-25'
  rubric: 0.23.0
  population: providers publishing a contract (contract_present earned), replayable exactly
  providers: 8977
  providers_unreplayable: 987
  providers_moved: 8927
  conditional_rows: excluded (they depend on what the API already does)
  composite_lift:
    median: 0.0
    p75: 0.2
    p90: 0.2
    max: 0.2
    mean_among_movers: 0.2
  agent_readiness_lift:
    median: 3.0
    p75: 3.1
    p90: 3.1
    max: 3.6
    mean_among_movers: 2.6
  facet_lift_median_among_movers:
    discoverability: 1.9
  composite_band_moves:
    thin -> developing: 40
    developing -> strong: 13
    emerging -> thin: 13
    strong -> exemplar: 3
  agent_readiness_band_moves:
    agent-aware -> agent-ready: 1098
    agent-ready -> agent-native: 114
  method: >-
    each provider's own kin/checks file, the vendor's maps at their stated credit, the scorer's composite
    formula; from -> to, nothing written