100 Thieves · Agentic Access

100 Thieves Agentic Access

x-agentic-access searched

100 Thieves exposes 13 API operations that an AI agent could call, of which 7 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: 6 read and 6 write.

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.

CompanyEsportsGamingApparelRetailE-CommerceEntertainmentMediaDirect To ConsumerAgentic Commerce
Operations: 13 Acting: 7 Human-in-the-loop: 1 Method: searched

By consequence

read 6 write 6

Source

Agentic Access

100-thieves-agentic-access.yml Raw ↑ 
generated: '2026-08-05'
method: searched
source:
- https://100thieves.com/agents.md
- https://100thieves.com/robots.txt
- https://100thieves.com/llms.txt
- mcp/100-thieves-tools-list.json
note: >-
  Unlike most entries in the catalog, this is not a generated governance starting
  point — 100 Thieves (via its Shopify UCP storefront) actually publishes agent access
  rules, so the contract below is transcribed from the provider's own statements and
  from the live tool schemas. derive-agentic-access.py was not run: it reads openapi/,
  and there is no OpenAPI here.
provider_published_policy:
  canonical_document: https://100thieves.com/agents.md
  mirrors: [https://100thieves.com/llms.txt]
  advertised_in: [robots.txt]
  rules:
  - id: buyer-approval-invariant
    statement: >-
      "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."
    source: https://100thieves.com/robots.txt
  - id: sanctioned-transaction-paths
    statement: Agents transacting on a buyer's behalf must use the UCP/MCP endpoint
      or the Shop skill; both require buyer approval before payment.
    paths: [https://100thieves.com/api/ucp/mcp, https://shop.app/SKILL.md]
  - id: rate-limits
    statement: The MCP endpoint is rate-limited per IP. Back off on 429 responses.
  - id: buyer-context
    statement: Pass context.address_country and context.currency for accurate pricing
      and availability.
  - id: crawling-permitted
    statement: Public product, collection, page, blog, policy, cart, and localized
      HTML is crawlable.
  - id: prefer-protocol-over-scraping
    statement: Agents should prefer the UCP/MCP endpoints or the Shop skill over screen-scraping
      or scripting the storefront directly.
  contact: bots@shopify.com
agent_identity:
  required: true
  mechanism: UCP agent profile
  field: meta.ucp-agent.profile
  format: URI
  enforcement: >-
    Enforced server-side. A tools/call without a resolvable profile URI returns JSON-RPC
    -32001 "UCP discovery failed" / invalid_profile_url with HTTP 422 — observed live.
    tools/list itself is anonymous.
summary:
  operations: 13
  by_action_class:
    connected: 6
    acting: 7
  by_consequence:
    read: 6
    write: 6
    financial: 1
  human_in_the_loop_required: 1
operations:
- tool: search_catalog
  action_class: connected
  consequence: read
  audit: optional
  human_in_the_loop: not-required
- tool: lookup_catalog
  action_class: connected
  consequence: read
  audit: optional
  human_in_the_loop: not-required
- tool: get_product
  action_class: connected
  consequence: read
  audit: optional
  human_in_the_loop: not-required
- tool: get_cart
  action_class: connected
  consequence: read
  audit: optional
  human_in_the_loop: not-required
- tool: get_checkout
  action_class: connected
  consequence: read
  audit: optional
  human_in_the_loop: not-required
- tool: get_order
  action_class: connected
  consequence: read
  audit: optional
  human_in_the_loop: not-required
  note: Reads buyer order data — treat the order id as a bearer capability.
- tool: create_cart
  action_class: acting
  consequence: write
  audit: required
  human_in_the_loop: not-required
- tool: update_cart
  action_class: acting
  consequence: write
  audit: required
  human_in_the_loop: not-required
- tool: cancel_cart
  action_class: acting
  consequence: write
  audit: required
  human_in_the_loop: not-required
- tool: create_checkout
  action_class: acting
  consequence: write
  audit: required
  human_in_the_loop: not-required
- tool: update_checkout
  action_class: acting
  consequence: write
  audit: required
  human_in_the_loop: not-required
  note: Carries shipping address — PII enters the call here.
- tool: cancel_checkout
  action_class: acting
  consequence: write
  audit: required
  human_in_the_loop: not-required
- tool: complete_checkout
  action_class: acting
  consequence: financial
  audit: required
  human_in_the_loop: required
  escalation: buyer-approval
  note: >-
    The one operation the provider explicitly fences. Payment is finalized here and
    the store's published policy forbids completing it without explicit, contemporaneous
    human approval. Agents that cannot obtain that approval are directed to route the
    purchase through Shop Pay via the Shop skill.
audience: null
x-evidence:
- url: https://100thieves.com/agents.md
  http_status: 200
- url: https://100thieves.com/robots.txt
  http_status: 200
- url: https://100thieves.com/api/ucp/mcp
  http_status: 422
  note: agent-profile gate observed on tools/call