Methodology

Four layers, normalized to a 0–100 score. Checks that don't apply to a product are dropped from the denominator — you're only measured on what you could have. Emerging checks are displayed but never scored; bonus checks can only raise a score.

contract v1.0.0

Discovery

weight 20 · Can agents find and recommend you?
  • Sitemap present & fresh
    required4 pts

    Resolved the sitemap from robots.txt or /sitemap.xml, validated XML, and checked lastmod freshness

  • Brand search discoverability
    required4 pts

    Ran clean brand-name web searches and checked whether your domain is cited in the results

  • Developer resource discoverability
    recommended3 pts

    Searched for your brand with developer-keyword suffixes and checked which official resources are cited

  • Homepage states what you are
    required3 pts

    An LLM read your homepage as an agent would and rated how confidently it could say what you do

  • OpenGraph / social metadata
    recommended2 pts

    Parsed homepage og:title / og:description / og:image and twitter:card meta tags

  • MCP registry listings
    recommended4 pts

    Queried the official MCP registry, Smithery, Glama, and PulseMCP for servers matching your domain

  • Agent skill published
    bonusemerging1 pts

    Checked /skill.md, /.well-known/skills/, and skills.sh for published agent skills

  • NLWeb endpoint
    emerging1 pts

    Probed /.well-known/nlweb.json and the /ask endpoint

Access

weight 30 · Can agents access your content and understand you?

Usability

weight 40 · Can agents integrate with and use you?
  • OpenAPI spec discoverable
    required5 pts

    Probed standard OpenAPI locations and docs links for a fetchable, parseable spec

  • OpenAPI validity & quality
    recommended3 pts

    Linted the spec: descriptions, operationIds, securitySchemes, servers

  • Docs discoverable
    required4 pts

    Followed homepage nav/footer links and probed /docs, /developers, docs.{domain}

  • Docs quality
    recommended3 pts

    An LLM rated your docs for clarity, completeness, runnable examples, and agent-friendliness

  • Quickstart / getting started
    recommended2 pts

    Looked for a quickstart/getting-started guide containing code blocks

  • MCP server discovery
    required5 pts

    Probed /.well-known/mcp/server-card.json, mcp.json, and docs mentions for an MCP endpoint

  • MCP handshake & tools list
    required4 pts

    Performed a streamable-HTTP initialize + tools/list against the MCP endpoint

  • MCP tool quality
    recommended3 pts

    Linted listed tools for descriptions, typed input schemas, and naming

  • npm SDK
    recommended2 pts

    Searched the npm registry for an official, domain-verified SDK package

  • PyPI SDK
    recommended2 pts

    Searched PyPI for an official, domain-verified SDK package

  • OAuth authorization server metadata
    recommended3 pts

    Fetched /.well-known/oauth-authorization-server (and openid-configuration fallback)

  • OAuth protected resource metadata
    recommended2 pts

    Fetched /.well-known/oauth-protected-resource

  • Auth documentation
    recommended2 pts

    Looked for /auth.md or an authentication docs page with code examples

  • agents.md
    recommended3 pts

    Fetched /agents.md and checked for substantive agent guidance

  • security.txt
    recommended1 pts

    Fetched /.well-known/security.txt and validated Contact + Expires

  • Accessibility basics
    required2 pts

    Static checks: lang attribute, title, alt coverage, labeled inputs, landmarks

  • Agent configs in public repo
    bonusemerging1 pts

    Checked your public GitHub org's main repos for AGENTS.md / .claude / .cursor configs

  • WebMCP
    emerging1 pts

    Looked for WebMCP script declarations on the homepage

Payments

weight 10 · Can agents transact with you?