generated: '2026-09-19'
method: derived
source: >-
  Derived by aligning the live MCP tools/list (8 tools, fetched anonymously 2026-09-19 and saved verbatim in
  mcp/bmcxiv-com-tools-list.json) with the provider's own OpenAPI 3.1 (openapi/bmcxiv-com-openapi.yml, 12
  operations). Both surfaces are first-party and unversioned relative to each other (both report 0.1.5), and the
  tool outputSchemas mirror the REST response schemas field for field, so bindings are high confidence.
purpose: >-
  Bind each MCP tool to the REST operation that backs it so an agent can read the tool's real contract from the
  OpenAPI, and make visible where the two surfaces diverge: two free MCP-only tools with no REST path, and the
  paid x402 call plus five discovery GETs that exist only over HTTP.
surfaces:
  rest_openapi: openapi/bmcxiv-com-openapi.yml        # 12 operations; source verbatim in openapi/_original/
  mcp: https://breach402.bmcxiv.com/mcp                # tools/list, initialize, resources/list all open (no auth)
  a2a: https://breach402.bmcxiv.com/a2a                # JSON-RPC; 5 skills on the card, not bound here (skills are prose, not tools)
  graphql: none
crosswalk:
- tool: create_owner_email_verification
  category: owner-verification
  rest: [createOwnerEmailVerification]
  binding: rest
  confidence: high
  note: tool outputSchema (enrollment_id, status, email, verification_expires_at, poll_token, status_url, verification_methods, next_step) equals EnrollmentCreated.
- tool: get_owner_email_verification_status
  category: owner-verification
  rest: [getOwnerEmailVerification]
  binding: rest
  confidence: high
  note: poll_token input = the REST bearer; output is EnrollmentStatus (scan_token present once verified).
- tool: verify_owner_email_code
  category: owner-verification
  rest: [verifyOwnerEmailCode]
  binding: rest
  confidence: high
- tool: revoke_owner_email_authorization
  category: owner-verification
  rest: [revokeOwnerEmailVerification]
  binding: rest
  confidence: high
  note: the only destructiveHint true tool; DELETE on REST.
- tool: prepare_paid_breach_check
  category: paid-check
  rest: [prepareOwnerBreachExposurePayment]
  binding: rest
  confidence: high
  note: >-
    Output is PreparedCheck plus already_paid and check — i.e. the tool returns the signed x402 HTTP request the
    agent must then perform itself against POST /v1/checks/run. The payment leg is not an MCP tool.
- tool: get_breach_report
  category: paid-check
  rest: [getOwnerBreachExposureReport]
  binding: rest
  confidence: high
  note: report_token input = the REST bearer; output is CheckStatus with the embedded ExposureReport.
mcp_only:
- tool: assess_owner_security_need
  reason: Free no-lookup decision aid (input scenario string); no REST path exists. Also an A2A skill (owner-security-recommendation).
- tool: preview_synthetic_breach_report
  reason: Free static synthetic report; no REST path exists. Also an A2A skill (synthetic-report-preview).
rest_only:
- capability: paid-check (payment leg)
  operations: [runOwnerBreachExposureCheck]
  reason: The x402 402-challenge / pay / retry exchange happens over plain HTTP with the ticket prepare_paid_breach_check returns; an MCP client cannot carry the payment header, so this call is deliberately not a tool.
- capability: discovery
  operations: [getOwnerProtectionSkill, getBreach402TrustContract, getAgentRecommendationGuide, getCyberExpertSkill, getSocialEngineeringProtectorSkill]
  reason: Served to MCP clients as resources (resources/list returns the same five URIs), not as tools.
- capability: transport
  operations: [mcpStreamableHttp]
  reason: The MCP endpoint itself, declared in the OpenAPI as POST /mcp.
coverage:
  tools_named: 8
  tools_bound: 6
  mcp_only: 2
  rest_operations_total: 12
  rest_operations_with_tool: 6
  rest_operations_exposed_as_mcp_resources: 5
  rest_only_functional: 1