{
  "fixes": {
    "slug": "mastercard-agent-pay",
    "name": "Mastercard Agent Pay",
    "listing": "https://www.anchorterminal.com/tools/mastercard-agent-pay",
    "markdown": "# Fix list: Mastercard Agent Pay\n\nFrom Anchor Terminal's listing at https://www.anchorterminal.com/tools/mastercard-agent-pay, the October 2026 research run, assessed 16 October 2026. Grade E, 42.5 out of 100.\n\nThis is everything the published grade says the listing lacks, the biggest possible gain to the total first. It comes from the reason given for each score, the checklist each category was scored against (https://www.anchorterminal.com/benchmark/#checklist), the provenance checks, the deductions, what we couldn't check and what the review panel asked for. A fix counts at the next check, once it's public.\n\nFor a coding agent working on Mastercard Agent Pay: work through the items below in the product, its docs and its public pages. Each category gives the reason for its score, with the points each checklist item earned, and the checklist itself, so the gap is the items that earned less than their points. Change the product, not the wording, and keep a note of what you changed and where it's published.\n\n## 1. Reliability, 40 out of 100, up to 12 more on the total\n\nWhy it scored 40: Protocol reading, as for AP2 and ACP, split as reference implementations 30, live deployments 25, spec stability 25 and test vectors 20. The one public implementation is the Verifiable Intent Python SDK, Apache-2.0 and installed from a clone (no PyPI package), with five example scripts. Mastercard publishes no Agent Pay SDK, and the merchant guides carry Web Bot Auth verification code only (12 of 30). Mastercard reports US launches, all European issuers enabled at network level, controlled live transactions with over 30 named European banks, and Australia's first authenticated agentic transactions with CBA and Westpac on 28 January 2026. Most are described as controlled trials (18 of 25). No public Agent Pay specification. The Acceptance Framework is linked to a product page we couldn't read, the early preview docs of 16 October 2025 now return 404, and Verifiable Intent is a v0.1 draft with a breaking field change on 17 April 2026 (5 of 25). Verifiable Intent has 326 test functions in CI, but no conformance vectors, and Agent Pay has no public sandbox (5 of 20). 40.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-reliability):\n\nHosted APIs, MCP servers, models and platforms.\n\n- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).\n- 0 to 30, the incident record for the last 90 days on that page. 30 for a clean record or trivial incidents only, 20 for minor incidents only, 10 for one major outage (an hour or more of a core API down, or errors across the board), 0 for several. 5 when there's no history we could read, and the note says so.\n- 15, rate limits documented with numbers.\n- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.\n- 10, an SLA published for any paid tier.\n- 10, the surface agents use is generally available, not beta or preview.\n\nLocal packages, SDKs, frameworks and stdio MCP servers.\n\n- 20, installs from an official package with supported runtimes stated.\n- 25, a public CI and test suite, passing on the default branch.\n- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).\n- 15, semver discipline and breaking changes called out in a changelog.\n- 15, version 1.0 or later, or declared stable.\n\nProtocols are read from their reference implementations, the public facilitators or servers, spec stability and test vectors.\n\n## 2. Agent ergonomics, 38 out of 100, up to 10.1 more on the total\n\nWhy it scored 38: Protocol reading. An implementer has to assemble Web Bot Auth (RFC 9421), the Agent User Identity object, agentic tokens, the Intent API and a three-layer SD-JWT chain from press releases, four merchant guides and a separate specification (8 of 25). Selective disclosure sends each party only its part of a Verifiable Intent chain (8 of 20). No documented Agent Pay errors. The guides tell merchants to reject or flag mismatches against an intent, which an agent sees only as a failed checkout (6 of 20). Web Bot Auth signatures carry a nonce and expiry, Verifiable Intent binds payment to one checkout by `checkout_hash` with a 5-minute L3 lifetime, and Intent orders carry an expiry (12 of 20). One SDK, Python, from git. The Agent Toolkit MCP server reads Mastercard documentation and doesn't call Agent Pay (4 of 15). 38.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-ergonomics):\n\n- 0 to 25, context cost. For MCP, the number and size of the tool definitions (25 for ten or fewer compact tools, 15 for 11 to 30, 5 for more than 30, plus up to 10 back for toolsets, dynamic loading or read-only subsets). For APIs, whether responses can be sized (field selection, limits, summaries).\n- 20, pagination, filtering and output-size controls.\n- 20, actionable, documented error responses, codes and messages an agent can recover from.\n- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.\n- 15, sensible defaults, few required parameters, and official SDKs in at least two languages.\n\nModels are read for tool use, structured output, prompt caching, context length, batch and SDKs. Frameworks for how much code and how many defaults a tool-calling agent with MCP needs.\n\n## 3. Payments \u0026 pricing, 25 out of 100, up to 9.4 more on the total\n\nWhy it scored 25: Protocol reading. Agent Pay moves a person's Mastercard through an agentic token after they authenticate, so an agent can't pay on its own. Agent Pay for Machines is announced for machine-speed payments with no documentation, and Mastercard's own text names no x402 or MPP support (10 of 40). No fee schedule for Agent Pay registration or tokens was found, though the specification and guides are free to read (5 of 20). Verifiable Intent and the merchant guides are free with no account. Agent Pay itself has no public sandbox or trial (10 of 20). Registration and token provisioning go through Mastercard onboarding (0). 25.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-payments):\n\nThe published rubric, also on the [x402 page](https://www.anchorterminal.com/x402/).\n\n- 40, a machine payment protocol (x402, MPP or L402) on the tool's own endpoints. 10 to 30 when it covers only some endpoints or only goes through a third party, and the note says which.\n- 20, per-call or per-unit pricing published without a login. 10 for public plan-only pricing, 0 for \"contact sales\" or prices behind a login.\n- 20, a free tier or trial that doesn't need a card.\n- 20, autonomous onboarding, meaning an agent can get access without a person signing up in a browser (keyless use, x402, a programmatic key API).\n\nPayment platforms and agent wallets rarely charge for their own API over a machine protocol, so the first line has steps for them, and the highest one that applies counts. 40 when x402, MPP or L402 runs on all their own endpoints, 30 when it runs on part of their own API, 25 when their merchants can accept one, 20 for running a facilitator, 15 for paying as a buyer, and 0 when the only protocol is their own. Merchant acceptance sits above a facilitator because the platform's own customers can charge agents through it, while a facilitator settles for sellers who wire up the protocol themselves. The counter-argument (a facilitator does more for the protocol as a whole) has a point. Each note says which step applied.\n\nOpen-source software you run yourself is scored on its hosted or paid option if it has one. A free, self-hosted package with nothing to buy gets 20, 20 and 20 for the last three lines, and 0 to 40 for the first only if it ships a payment protocol.\n\n## 4. Schema \u0026 documentation, 47 out of 100, up to 8.6 more on the total\n\nWhy it scored 47: No public Agent Pay API contract. The portal's Agent Pay API reference is a template whose Swagger link is a petstore file that answers 401. Verifiable Intent defines its credentials in Markdown with JSON examples and Python models, with no JSON Schema (8 of 25). developer.mastercard.com serves llms.txt and Markdown for every page, but Agent Pay has no entry in it. Verifiable Intent's site has no llms.txt, though its Markdown sources are in the repository (7 of 10). The Verifiable Intent specification runs to 4,077 lines with MUST and SHOULD rules, and the merchant guides describe the Agent User Identity object, the Intent object and the Web Bot Auth tags in prose and tables (12 of 20). Verifiable Intent constraints are typed, such as `mandate.payment.amount_range` with currency and integer minor units. The Intent fields in the guides carry no types (9 of 15). Examples in the specification and guides. No Agent Pay error codes (7 of 15). Verifiable Intent keeps a CHANGELOG.md with one entry, and its April breaking change isn't in it. Agent Pay has no changelog (4 of 15). 47.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-schema):\n\nAPIs and MCP servers.\n\n- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).\n- 10, llms.txt or Markdown docs served for agents.\n- 0 to 20, descriptions that say what a tool is for, when to use it and when not to, read from the tool definitions in the source or the API reference.\n- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.\n- 0 to 15, examples and documented error responses.\n- 15, versioning and a public changelog.\n\nModels are read from the API reference, the OpenAPI file, llms.txt, the structured-output and tool-use docs and the model cards. Frameworks from docs a model can follow, typed interfaces, examples and the API reference.\n\n## 5. Maintenance \u0026 community, 23 out of 100, up to 6.7 more on the total\n\nWhy it scored 23: The last dated change is Agent Pay for Machines on 10 June 2026, 128 days before this check (10). No three dated releases in the last 90 days. The last Verifiable Intent commit is 20 April 2026 (0). No public Agent Pay changelog or issue tracker. Verifiable Intent's issues couldn't be read because the GitHub API was refused by our session. Mastercard Developers has a support route (5 of 15 for a closed service). No official Agent Pay SDK, and Verifiable Intent's SDK isn't on PyPI (3 of 15). Verifiable Intent CI runs ruff and pytest on three operating systems and Python 3.10 to 3.13, with actions pinned to tags (5 of 10). 23.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-maintenance):\n\n- 0 to 30, time since the last release, or the last published model or API change for a closed service. 30 within 30 days, 20 within 90, 10 within 180, 0 older.\n- 20, at least three releases or dated changelog entries in the last 90 days.\n- 0 to 25, responsiveness. Issues and pull requests answered on GitHub (the open issues and how recent the replies are). For closed services, a public changelog and a support or community channel that answers, 0 to 15.\n- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).\n- 10, package health, current dependencies and CI.\n\nModels are read for deprecation notice periods and model churn rather than release counts.\n\n## 6. Transparency \u0026 trust, 36 out of 100, up to 5.6 more on the total\n\nMade of editorial 31, provenance 41.\n\nWhy it scored 36: Agent Pay is closed and its programme rules weren't found in public. Verifiable Intent is Apache-2.0 under a DCO (15 of 30). No Agent Pay data handling statement was found. Verifiable Intent's selective disclosure limits what each party sees, and Mastercard's privacy notice couldn't be read (8 of 30). The early preview docs were removed with no notice in the release history, and Verifiable Intent's breaking change has no changelog entry (3 of 20). No sub-processors or data locations were found for Agent Pay (5 of 20). 31.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-transparency):\n\n- 0 to 30, source availability and licence clarity. 30 for open source under an OSI licence, 15 for closed with clear terms, 0 for unclear terms.\n- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).\n- 0 to 20, a deprecation policy or notices with dates.\n- 0 to 20, telemetry disclosed with an opt-out (local software), or subprocessors and data locations disclosed (hosted).\n\nThe other half of Transparency and trust is the provenance score, computed from checked facts (below). The category score is the mean of the two.\n\nProvenance checks not met in full (half of this category, computed from checked facts):\n\n- Legal entity named: not found (0 of 20)\n- Status page: not found (0 of 10)\n- Changelog: not found (0 of 10)\n- security.txt: could not be fetched (0 of 10)\n\n## 7. Security \u0026 auth, 71 out of 100, up to 5.1 more on the total\n\nWhy it scored 71: Agentic tokens keep the card number from the agent, may be single-use, merchant-specific or time-limited, and carry a one-time cryptogram. Agents sign requests with keys registered in Mastercard's directory. Verifiable Intent key-binds each layer through `cnf`, but v0.1 defines no credential revocation, and no token revocation for Agent Pay is documented publicly (22 of 30). Mastercard says consumers control what the agent may buy. European live transactions used passkeys, and Verifiable Intent constrains amount range, payees, merchants and line items, with an Immediate mode for user-confirmed purchases (16 of 20). The Verifiable Intent security model bounds a compromised or malicious agent by the user's constraints. Neither it nor the guides address prompt injection by name (10 of 15). Intent records and Verifiable Intent chains give issuers and merchants an audit trail for disputes (11 of 15). Mastercard runs a public bug bounty on Bugcrowd and Verifiable Intent's SECURITY.md takes reports by email or GitHub private reporting within 10 working days. security.txt on mastercard.com couldn't be read and on developer.mastercard.com redirects to a login. No advisories (12 of 20). 71.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-security):\n\n- 0 to 30, the credential model. 30 for OAuth 2.1 with scopes, or scoped and revocable keys with rotation. 20 for plain revocable API keys. 10 for one all-powerful key. 10 off when a secret can travel in a URL query string as a documented option.\n- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.\n- 0 to 15, prompt-injection posture where the tool returns untrusted content (documented mitigations or guidance). A tool that returns no untrusted content gets 10.\n- 0 to 15, audit logs or per-call visibility for the operator.\n- 0 to 20, a security programme. security.txt or a disclosure policy, a bug bounty, SOC 2 or ISO 27001, advisories handled in public.\n\nModels are read for retention, whether API data trains models (and whether that's off by default), zero-retention options and certifications. Frameworks for telemetry defaults, approval hooks, guardrails and sandboxing.\n\n## What we couldn't check\n\nWhat we couldn't read counted as absent. Publishing it on a page a plain HTTP fetch can read (not only in a browser) lets the next check count it.\n\n- unchecked: the Agent Pay product page and the Agent Pay Acceptance Framework it links, which the developer guides cite for the complete Intent API specification. WebFetch permission for www.mastercard.com product pages wasn't answered, and the site answers curl with 403\n- unchecked: the Agent Pay for Machines product page, for the same reason\n- unchecked: the legal entity, terms of use and privacy notice behind mastercard.com, and security.txt on www.mastercard.com\n- unchecked: whether the developer portal's Agent Pay docs exist behind a login. The Markdown paths return 404, the public sitemap lists only two template product pages, and unknown paths redirect to a login\n- unchecked: agentpay-key-directory.mastercard.com, named in the Level 1 guide as the Agent Pay key directory. Our proxy returned 502 for it\n- unchecked: GitHub stars, issues and pull requests for agent-intent/verifiable-intent, because the GitHub API was refused by our session\n- Whether Agent Pay for Machines supports x402 or MPP. Mastercard's own text doesn't say so. The founder's lead said it does, from partner quotes by Coinbase (x402) and Tempo (MPP compatibility work)\n- Whether Agent Pay has dispute or purchase protection rules of its own. The pages read say intent records help resolve disputes and that existing chargeback rights apply\n- Whether the Intent API in the Level 3 guide is live and whether Verifiable Intent is now integrated into it, as Mastercard said it would be in March 2026\n- Whether the removal of the October 2025 preview docs while still linked from Mastercard's framework post should carry a deduction. We took none, since no endpoint is involved\n\n## Weaknesses\n\n- No public Agent Pay API reference, SDK or sandbox. The portal's Agent Pay API page is a template pointing at a petstore spec\n- Agent registration, certification and token provisioning need Mastercard onboarding, with no published process or fees\n- The early preview docs of 16 October 2025 were removed with no notice in the release history, and Mastercard's own post still links to them\n- Verifiable Intent is a v0.1 draft with no revocation protocol, no PyPI package and no commit since 20 April 2026\n- No Agent Pay dispute or purchase protection rules were found in the Mastercard pages read\n\n## What costs an agent a turn today\n\nThe notes we give agents before they call it. Each one is a workaround an agent shouldn't need.\n\n- Don't plan to call Agent Pay directly. Registration and agentic tokens come through Mastercard onboarding, so hand card purchases to the person or a registered agent platform\n- If you are registered, sign every merchant request with Web Bot Auth, tag `Agent-pay-auth`, and `Signature-Agent` set to Mastercard's key directory\n- Submit the agentic token in the normal card fields and never ask for or store a card number\n- For delegated purchases, ask the user for Verifiable Intent constraints with an amount range, allowed merchants and the shortest lifetime that fits\n- Read the merchant guides on developer.mastercard.com as Markdown. The Agent Pay docs section there has no public content\n\n## What the review panel asked for\n\n- Publish the registration process\n- Open a public sandbox\n- an Agent Pay changelog\n- notice before removing docs\n- Publish registration and token fees\n- Open a trial sandbox\n- add revocation to Verifiable Intent\n- document token revocation\n\n## When it's done\n\nSend what changed and where it's published as a dispute (https://www.anchorterminal.com/builders/#disputes, or `POST https://www.anchorterminal.com/api/v1/contact` with `\"kind\": \"dispute\"`). Disputes are answered in public, and the listing is checked again by the same checklist. Paying for an audit or a listing claim changes nothing here.\n",
    "grade": "E",
    "score": 42.5,
    "assessed": "2026-10-16",
    "run": "October 2026 research run",
    "categories": [
      {
        "key": "reliability",
        "name": "Reliability",
        "score": 40,
        "maxGain": 12,
        "reason": "Protocol reading, as for AP2 and ACP, split as reference implementations 30, live deployments 25, spec stability 25 and test vectors 20. The one public implementation is the Verifiable Intent Python SDK, Apache-2.0 and installed from a clone (no PyPI package), with five example scripts. Mastercard publishes no Agent Pay SDK, and the merchant guides carry Web Bot Auth verification code only (12 of 30). Mastercard reports US launches, all European issuers enabled at network level, controlled live transactions with over 30 named European banks, and Australia's first authenticated agentic transactions with CBA and Westpac on 28 January 2026. Most are described as controlled trials (18 of 25). No public Agent Pay specification. The Acceptance Framework is linked to a product page we couldn't read, the early preview docs of 16 October 2025 now return 404, and Verifiable Intent is a v0.1 draft with a breaking field change on 17 April 2026 (5 of 25). Verifiable Intent has 326 test functions in CI, but no conformance vectors, and Agent Pay has no public sandbox (5 of 20). 40.",
        "checklist": [
          "Hosted APIs, MCP servers, models and platforms.",
          "- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).\n- 0 to 30, the incident record for the last 90 days on that page. 30 for a clean record or trivial incidents only, 20 for minor incidents only, 10 for one major outage (an hour or more of a core API down, or errors across the board), 0 for several. 5 when there's no history we could read, and the note says so.\n- 15, rate limits documented with numbers.\n- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.\n- 10, an SLA published for any paid tier.\n- 10, the surface agents use is generally available, not beta or preview.",
          "Local packages, SDKs, frameworks and stdio MCP servers.",
          "- 20, installs from an official package with supported runtimes stated.\n- 25, a public CI and test suite, passing on the default branch.\n- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).\n- 15, semver discipline and breaking changes called out in a changelog.\n- 15, version 1.0 or later, or declared stable.",
          "Protocols are read from their reference implementations, the public facilitators or servers, spec stability and test vectors."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-reliability"
      },
      {
        "key": "ergonomics",
        "name": "Agent ergonomics",
        "score": 38,
        "maxGain": 10.1,
        "reason": "Protocol reading. An implementer has to assemble Web Bot Auth (RFC 9421), the Agent User Identity object, agentic tokens, the Intent API and a three-layer SD-JWT chain from press releases, four merchant guides and a separate specification (8 of 25). Selective disclosure sends each party only its part of a Verifiable Intent chain (8 of 20). No documented Agent Pay errors. The guides tell merchants to reject or flag mismatches against an intent, which an agent sees only as a failed checkout (6 of 20). Web Bot Auth signatures carry a nonce and expiry, Verifiable Intent binds payment to one checkout by `checkout_hash` with a 5-minute L3 lifetime, and Intent orders carry an expiry (12 of 20). One SDK, Python, from git. The Agent Toolkit MCP server reads Mastercard documentation and doesn't call Agent Pay (4 of 15). 38.",
        "checklist": [
          "- 0 to 25, context cost. For MCP, the number and size of the tool definitions (25 for ten or fewer compact tools, 15 for 11 to 30, 5 for more than 30, plus up to 10 back for toolsets, dynamic loading or read-only subsets). For APIs, whether responses can be sized (field selection, limits, summaries).\n- 20, pagination, filtering and output-size controls.\n- 20, actionable, documented error responses, codes and messages an agent can recover from.\n- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.\n- 15, sensible defaults, few required parameters, and official SDKs in at least two languages.",
          "Models are read for tool use, structured output, prompt caching, context length, batch and SDKs. Frameworks for how much code and how many defaults a tool-calling agent with MCP needs."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-ergonomics"
      },
      {
        "key": "payments",
        "name": "Payments \u0026 pricing",
        "score": 25,
        "maxGain": 9.4,
        "reason": "Protocol reading. Agent Pay moves a person's Mastercard through an agentic token after they authenticate, so an agent can't pay on its own. Agent Pay for Machines is announced for machine-speed payments with no documentation, and Mastercard's own text names no x402 or MPP support (10 of 40). No fee schedule for Agent Pay registration or tokens was found, though the specification and guides are free to read (5 of 20). Verifiable Intent and the merchant guides are free with no account. Agent Pay itself has no public sandbox or trial (10 of 20). Registration and token provisioning go through Mastercard onboarding (0). 25.",
        "checklist": [
          "The published rubric, also on the [x402 page](https://www.anchorterminal.com/x402/).",
          "- 40, a machine payment protocol (x402, MPP or L402) on the tool's own endpoints. 10 to 30 when it covers only some endpoints or only goes through a third party, and the note says which.\n- 20, per-call or per-unit pricing published without a login. 10 for public plan-only pricing, 0 for \"contact sales\" or prices behind a login.\n- 20, a free tier or trial that doesn't need a card.\n- 20, autonomous onboarding, meaning an agent can get access without a person signing up in a browser (keyless use, x402, a programmatic key API).",
          "Payment platforms and agent wallets rarely charge for their own API over a machine protocol, so the first line has steps for them, and the highest one that applies counts. 40 when x402, MPP or L402 runs on all their own endpoints, 30 when it runs on part of their own API, 25 when their merchants can accept one, 20 for running a facilitator, 15 for paying as a buyer, and 0 when the only protocol is their own. Merchant acceptance sits above a facilitator because the platform's own customers can charge agents through it, while a facilitator settles for sellers who wire up the protocol themselves. The counter-argument (a facilitator does more for the protocol as a whole) has a point. Each note says which step applied.",
          "Open-source software you run yourself is scored on its hosted or paid option if it has one. A free, self-hosted package with nothing to buy gets 20, 20 and 20 for the last three lines, and 0 to 40 for the first only if it ships a payment protocol."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-payments"
      },
      {
        "key": "schema",
        "name": "Schema \u0026 documentation",
        "score": 47,
        "maxGain": 8.6,
        "reason": "No public Agent Pay API contract. The portal's Agent Pay API reference is a template whose Swagger link is a petstore file that answers 401. Verifiable Intent defines its credentials in Markdown with JSON examples and Python models, with no JSON Schema (8 of 25). developer.mastercard.com serves llms.txt and Markdown for every page, but Agent Pay has no entry in it. Verifiable Intent's site has no llms.txt, though its Markdown sources are in the repository (7 of 10). The Verifiable Intent specification runs to 4,077 lines with MUST and SHOULD rules, and the merchant guides describe the Agent User Identity object, the Intent object and the Web Bot Auth tags in prose and tables (12 of 20). Verifiable Intent constraints are typed, such as `mandate.payment.amount_range` with currency and integer minor units. The Intent fields in the guides carry no types (9 of 15). Examples in the specification and guides. No Agent Pay error codes (7 of 15). Verifiable Intent keeps a CHANGELOG.md with one entry, and its April breaking change isn't in it. Agent Pay has no changelog (4 of 15). 47.",
        "checklist": [
          "APIs and MCP servers.",
          "- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).\n- 10, llms.txt or Markdown docs served for agents.\n- 0 to 20, descriptions that say what a tool is for, when to use it and when not to, read from the tool definitions in the source or the API reference.\n- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.\n- 0 to 15, examples and documented error responses.\n- 15, versioning and a public changelog.",
          "Models are read from the API reference, the OpenAPI file, llms.txt, the structured-output and tool-use docs and the model cards. Frameworks from docs a model can follow, typed interfaces, examples and the API reference."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-schema"
      },
      {
        "key": "maintenance",
        "name": "Maintenance \u0026 community",
        "score": 23,
        "maxGain": 6.7,
        "reason": "The last dated change is Agent Pay for Machines on 10 June 2026, 128 days before this check (10). No three dated releases in the last 90 days. The last Verifiable Intent commit is 20 April 2026 (0). No public Agent Pay changelog or issue tracker. Verifiable Intent's issues couldn't be read because the GitHub API was refused by our session. Mastercard Developers has a support route (5 of 15 for a closed service). No official Agent Pay SDK, and Verifiable Intent's SDK isn't on PyPI (3 of 15). Verifiable Intent CI runs ruff and pytest on three operating systems and Python 3.10 to 3.13, with actions pinned to tags (5 of 10). 23.",
        "checklist": [
          "- 0 to 30, time since the last release, or the last published model or API change for a closed service. 30 within 30 days, 20 within 90, 10 within 180, 0 older.\n- 20, at least three releases or dated changelog entries in the last 90 days.\n- 0 to 25, responsiveness. Issues and pull requests answered on GitHub (the open issues and how recent the replies are). For closed services, a public changelog and a support or community channel that answers, 0 to 15.\n- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).\n- 10, package health, current dependencies and CI.",
          "Models are read for deprecation notice periods and model churn rather than release counts."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-maintenance"
      },
      {
        "key": "transparency",
        "name": "Transparency \u0026 trust",
        "score": 36,
        "maxGain": 5.6,
        "reason": "Agent Pay is closed and its programme rules weren't found in public. Verifiable Intent is Apache-2.0 under a DCO (15 of 30). No Agent Pay data handling statement was found. Verifiable Intent's selective disclosure limits what each party sees, and Mastercard's privacy notice couldn't be read (8 of 30). The early preview docs were removed with no notice in the release history, and Verifiable Intent's breaking change has no changelog entry (3 of 20). No sub-processors or data locations were found for Agent Pay (5 of 20). 31.",
        "blend": "editorial 31, provenance 41",
        "checklist": [
          "- 0 to 30, source availability and licence clarity. 30 for open source under an OSI licence, 15 for closed with clear terms, 0 for unclear terms.\n- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).\n- 0 to 20, a deprecation policy or notices with dates.\n- 0 to 20, telemetry disclosed with an opt-out (local software), or subprocessors and data locations disclosed (hosted).",
          "The other half of Transparency and trust is the provenance score, computed from checked facts (below). The category score is the mean of the two."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-transparency"
      },
      {
        "key": "security",
        "name": "Security \u0026 auth",
        "score": 71,
        "maxGain": 5.1,
        "reason": "Agentic tokens keep the card number from the agent, may be single-use, merchant-specific or time-limited, and carry a one-time cryptogram. Agents sign requests with keys registered in Mastercard's directory. Verifiable Intent key-binds each layer through `cnf`, but v0.1 defines no credential revocation, and no token revocation for Agent Pay is documented publicly (22 of 30). Mastercard says consumers control what the agent may buy. European live transactions used passkeys, and Verifiable Intent constrains amount range, payees, merchants and line items, with an Immediate mode for user-confirmed purchases (16 of 20). The Verifiable Intent security model bounds a compromised or malicious agent by the user's constraints. Neither it nor the guides address prompt injection by name (10 of 15). Intent records and Verifiable Intent chains give issuers and merchants an audit trail for disputes (11 of 15). Mastercard runs a public bug bounty on Bugcrowd and Verifiable Intent's SECURITY.md takes reports by email or GitHub private reporting within 10 working days. security.txt on mastercard.com couldn't be read and on developer.mastercard.com redirects to a login. No advisories (12 of 20). 71.",
        "checklist": [
          "- 0 to 30, the credential model. 30 for OAuth 2.1 with scopes, or scoped and revocable keys with rotation. 20 for plain revocable API keys. 10 for one all-powerful key. 10 off when a secret can travel in a URL query string as a documented option.\n- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.\n- 0 to 15, prompt-injection posture where the tool returns untrusted content (documented mitigations or guidance). A tool that returns no untrusted content gets 10.\n- 0 to 15, audit logs or per-call visibility for the operator.\n- 0 to 20, a security programme. security.txt or a disclosure policy, a bug bounty, SOC 2 or ISO 27001, advisories handled in public.",
          "Models are read for retention, whether API data trains models (and whether that's off by default), zero-retention options and certifications. Frameworks for telemetry defaults, approval hooks, guardrails and sandboxing."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-security"
      }
    ],
    "provenance": [
      {
        "label": "Legal entity named",
        "value": "not found",
        "points": 0,
        "max": 20
      },
      {
        "label": "Status page",
        "value": "not found",
        "points": 0,
        "max": 10
      },
      {
        "label": "Changelog",
        "value": "not found",
        "points": 0,
        "max": 10
      },
      {
        "label": "security.txt",
        "value": "could not be fetched",
        "points": 0,
        "max": 10
      }
    ],
    "unchecked": [
      "unchecked: the Agent Pay product page and the Agent Pay Acceptance Framework it links, which the developer guides cite for the complete Intent API specification. WebFetch permission for www.mastercard.com product pages wasn't answered, and the site answers curl with 403",
      "unchecked: the Agent Pay for Machines product page, for the same reason",
      "unchecked: the legal entity, terms of use and privacy notice behind mastercard.com, and security.txt on www.mastercard.com",
      "unchecked: whether the developer portal's Agent Pay docs exist behind a login. The Markdown paths return 404, the public sitemap lists only two template product pages, and unknown paths redirect to a login",
      "unchecked: agentpay-key-directory.mastercard.com, named in the Level 1 guide as the Agent Pay key directory. Our proxy returned 502 for it",
      "unchecked: GitHub stars, issues and pull requests for agent-intent/verifiable-intent, because the GitHub API was refused by our session",
      "Whether Agent Pay for Machines supports x402 or MPP. Mastercard's own text doesn't say so. The founder's lead said it does, from partner quotes by Coinbase (x402) and Tempo (MPP compatibility work)",
      "Whether Agent Pay has dispute or purchase protection rules of its own. The pages read say intent records help resolve disputes and that existing chargeback rights apply",
      "Whether the Intent API in the Level 3 guide is live and whether Verifiable Intent is now integrated into it, as Mastercard said it would be in March 2026",
      "Whether the removal of the October 2025 preview docs while still linked from Mastercard's framework post should carry a deduction. We took none, since no endpoint is involved"
    ],
    "weaknesses": [
      "No public Agent Pay API reference, SDK or sandbox. The portal's Agent Pay API page is a template pointing at a petstore spec",
      "Agent registration, certification and token provisioning need Mastercard onboarding, with no published process or fees",
      "The early preview docs of 16 October 2025 were removed with no notice in the release history, and Mastercard's own post still links to them",
      "Verifiable Intent is a v0.1 draft with no revocation protocol, no PyPI package and no commit since 20 April 2026",
      "No Agent Pay dispute or purchase protection rules were found in the Mastercard pages read"
    ],
    "agentNotes": [
      "Don't plan to call Agent Pay directly. Registration and agentic tokens come through Mastercard onboarding, so hand card purchases to the person or a registered agent platform",
      "If you are registered, sign every merchant request with Web Bot Auth, tag `Agent-pay-auth`, and `Signature-Agent` set to Mastercard's key directory",
      "Submit the agentic token in the normal card fields and never ask for or store a card number",
      "For delegated purchases, ask the user for Verifiable Intent constraints with an amount range, allowed merchants and the shortest lifetime that fits",
      "Read the merchant guides on developer.mastercard.com as Markdown. The Agent Pay docs section there has no public content"
    ],
    "requests": [
      {
        "text": "Publish the registration process",
        "reviews": 1
      },
      {
        "text": "Open a public sandbox",
        "reviews": 1
      },
      {
        "text": "an Agent Pay changelog",
        "reviews": 1
      },
      {
        "text": "notice before removing docs",
        "reviews": 1
      },
      {
        "text": "Publish registration and token fees",
        "reviews": 1
      },
      {
        "text": "Open a trial sandbox",
        "reviews": 1
      },
      {
        "text": "add revocation to Verifiable Intent",
        "reviews": 1
      },
      {
        "text": "document token revocation",
        "reviews": 1
      }
    ],
    "recheck": "https://www.anchorterminal.com/builders/#disputes"
  },
  "meta": {
    "attribution": "Anchor Terminal (https://www.anchorterminal.com)",
    "attributionRequired": true,
    "credit": "CC BY 4.0, credit required: when you show, quote or pass on anything from here (to a user in an answer, in a product, a report or a dataset), say it's from Anchor Terminal (anchorterminal.com) and link the page or API address you took it from. Reviews also credit their author.",
    "docs": "https://www.anchorterminal.com/docs/",
    "generatedAt": "2026-10-16",
    "license": "CC-BY-4.0",
    "method": "https://www.anchorterminal.com/benchmark/",
    "methodology": "0.4",
    "openapi": "https://www.anchorterminal.com/openapi.json",
    "preview": false,
    "run": "2026-10-01",
    "runLabel": "October 2026 research run",
    "terms": "https://www.anchorterminal.com/terms/#data"
  }
}