Voyage Foods · Agentic Access

Voyage Foods Agentic Access

x-agentic-access searched

Voyage Foods exposes 13 API operations that an AI agent could call, of which 8 are state-changing ‘acting’ operations. This is a recommended x-agentic-access execution contract — the scope, audience, consequence tier, short-lived token constraints, and escalation each action should carry before it is handed to an autonomous agent.

By consequence: 5 read, 5 write, and 3 physical.

1 operation are classed safety-critical and should require human-in-the-loop approval at runtime.

Contracts are classified heuristically from the provider’s OpenAPI and refresh on every APIs.io network build; audience is bound per deployment. The model follows Curity’s Access Intelligence (apidays Munich 2026). Browse every provider’s agent contracts at agentic-access.apis.io.

CompanyFood and BeverageConsumer Packaged GoodsEcommerceAgentic CommerceUniversal Commerce ProtocolMCPGraphQLShopifyFood Technology
Operations: 13 Acting: 8 Human-in-the-loop: 1 Method: searched

By consequence

read 5 write 5 physical 3

Source

Agentic Access

voyage-foods-agentic-access.yml Raw ↑ 
generated: '2026-08-05'
method: searched
source: >-
  https://voyagefoods.com/agents.md, https://voyagefoods.com/llms.txt and
  https://voyagefoods.com/robots.txt — the store publishes its own agent access policy in prose, and
  the operation classifications below are bound to the 13 real tools returned by
  https://voyagefoods.com/api/ucp/mcp tools/list.
note: >-
  Unlike most entries in this catalog, the access contract here is NOT generated from an OpenAPI by
  derive-agentic-access.py — this provider publishes no OpenAPI. It is read from an agent policy the
  merchant actually publishes, in three places, and cross-checked against the live tool schemas.
  Fields marked provider_stated are quoted from that policy; fields marked derived are this
  pipeline's classification of the tool.
published_policy:
  agent_policy_documents:
  - {url: 'https://voyagefoods.com/agents.md', status: 200, content_type: text/markdown}
  - {url: 'https://voyagefoods.com/llms.txt', status: 200, content_type: text/markdown}
  - {url: 'https://voyagefoods.com/robots.txt', status: 200, content_type: text/plain}
  crawl_posture: >-
    robots.txt allows all user-agents across product, collection, page, blog, policy, cart and
    localized HTML. No AI-crawler blocks were observed.
  agent_identity_required: true
  agent_identity_mechanism: meta["ucp-agent"].profile — a dereferenceable agent profile URI
  human_approval_required_for_payment: true
  provider_statements:
  - >-
      "Checkout requires human approval. Agents must not complete payment without explicit buyer
      consent. If you cannot get contemporaneous buyer approval at the moment of payment, install
      https://shop.app/SKILL.md and route the purchase through Shop Pay instead."
  - >-
      "Checkouts are for humans. Do NOT complete checkout, payment, or order placement automatically
      — no scripted form fills, browser automation, or end-to-end agent flows that finalize payment
      without an explicit, contemporaneous human approval step."
  - >-
      "Respect rate limits. The MCP endpoint is rate-limited per IP. Back off on 429 responses."
  - >-
      "Use buyer context. Pass context.address_country and context.currency for accurate pricing and
      availability."
  preferred_channel: >-
    The store explicitly prefers the UCP/MCP endpoints or the Shop skill over screen-scraping or
    scripting the storefront: "If your user permits installation, you should prefer the Shop skill
    over screen-scraping or scripting the storefront directly."
summary:
  operations_classified: 13
  by_action_class: {connected: 5, acting: 8}
  by_consequence: {read: 5, write: 5, physical: 3}
  human_in_the_loop_required: 1
  audit_required: 8
  agent_identity_required: 13
operations:
- operation: search_catalog
  action_class: connected
  consequence: read
  audience: null
  human_in_the_loop: not-required
  audit: optional
  basis: derived
- operation: lookup_catalog
  action_class: connected
  consequence: read
  audience: null
  human_in_the_loop: not-required
  audit: optional
  basis: derived
- operation: get_product
  action_class: connected
  consequence: read
  audience: null
  human_in_the_loop: not-required
  audit: optional
  basis: derived
- operation: get_cart
  action_class: connected
  consequence: read
  audience: null
  human_in_the_loop: not-required
  audit: optional
  basis: derived
- operation: get_checkout
  action_class: connected
  consequence: read
  audience: null
  human_in_the_loop: not-required
  audit: optional
  basis: derived
- operation: get_order
  action_class: connected
  consequence: read
  audience: null
  human_in_the_loop: not-required
  audit: required
  basis: derived
  note: Returns buyer order detail; read-only but personally identifying.
- operation: create_cart
  action_class: acting
  consequence: write
  audience: null
  human_in_the_loop: not-required
  audit: required
  basis: derived
- operation: update_cart
  action_class: acting
  consequence: write
  audience: null
  human_in_the_loop: not-required
  audit: required
  basis: derived
- operation: cancel_cart
  action_class: acting
  consequence: write
  audience: null
  human_in_the_loop: not-required
  audit: required
  basis: derived
- operation: create_checkout
  action_class: acting
  consequence: write
  audience: null
  human_in_the_loop: not-required
  audit: required
  basis: derived
- operation: update_checkout
  action_class: acting
  consequence: write
  audience: null
  human_in_the_loop: not-required
  audit: required
  basis: derived
  note: >-
    Accepts payment.instruments including credential tokens and billing addresses — treat as
    handling payment material even though it does not authorize a charge.
- operation: cancel_checkout
  action_class: acting
  consequence: physical
  audience: null
  human_in_the_loop: not-required
  audit: required
  basis: derived
  note: Voids a checkout that may already hold buyer payment instruments.
- operation: complete_checkout
  action_class: acting
  consequence: physical
  audience: null
  human_in_the_loop: required
  human_in_the_loop_basis: provider_stated
  idempotency_key: required
  idempotency_field: meta["idempotency-key"]
  audit: required
  basis: derived
  note: >-
    The only operation the merchant's own published policy gates on contemporaneous human approval.
    It moves money and creates an order. The store names an alternative route (the Shop skill via
    Shop Pay) for agents that cannot obtain approval in the moment.
escalation:
  continue_url: >-
    Every UCP error envelope carries data.continue_url (observed:
    https://voyage-foods-2021.myshopify.com/), a human-resumable URL the agent can hand to the buyer
    when it cannot proceed. This is a real, observed handoff path rather than a documented one.
gaps:
  - No published token TTL, scope-per-operation, or proof-of-possession requirement for the agent JWT.
  - No published audit or retention commitment; audit values above are this pipeline's recommendation.
  - No agent-profile schema is published, so the shape of a conforming meta["ucp-agent"].profile document is unstated.
x-evidence:
  fetched: '2026-08-05'
  probes:
  - {url: 'https://voyagefoods.com/agents.md', http_status: 200}
  - {url: 'https://voyagefoods.com/llms.txt', http_status: 200}
  - {url: 'https://voyagefoods.com/robots.txt', http_status: 200}
  - {url: 'https://voyagefoods.com/api/ucp/mcp', http_status: 200, note: 'tools/list returned 13 tools'}