openapi: 3.1.0
info:
  title: HumanMirror Sanitize Shield
  version: 2.0.0
  description: 'Quarantine external context, detect indirect prompt injection and issue a signed Cleanliness Proof. Three
    test calls per 24 hours, then x402 Base/USDC. Granular unit surface: Canonical bundle: https://humanmirror.fr/api/v1/agent/security-intelligence/'
servers:
- url: https://humanmirror.fr
paths:
  /api/v1/sanitize/shield/:
    get:
      operationId: discoverSanitizeShield
      responses:
        '200':
          description: Machine-readable service discovery
      x-humanmirror-paid-price-usdc: '0.001'
      x-humanmirror-canonical-bundle: https://humanmirror.fr/api/v1/agent/security-intelligence/
      x-humanmirror-pricing-tier: granular_unit_0_001
    post:
      operationId: sanitizeShield
      summary: Analyze and quarantine untrusted agent context
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/ShieldRequest'
            examples:
              structured:
                value:
                  payload:
                    source: web
                    content: Release notes for version 2.0
                  context:
                    trust_level: untrusted
        responses:
          '200':
            description: Analysis completed
            content:
              application/json:
                schema:
                  $ref: '#/components/schemas/ShieldResponse'
          '400':
            description: Invalid or oversized payload
          '402':
            description: Free quota exhausted; x402 payment required
            headers:
              PAYMENT-REQUIRED:
                schema:
                  type: string
                description: Base64url x402 V2 payment requirements
          '503':
            description: Quota, proof-signing or payment service unavailable
      x-humanmirror-paid-price-usdc: '0.001'
      x-humanmirror-canonical-bundle: https://humanmirror.fr/api/v1/agent/security-intelligence/
      x-humanmirror-pricing-tier: granular_unit_0_001
  components:
    schemas:
      ShieldRequest:
        type: object
        required:
        - payload
        properties:
          payload: {}
          context:
            type: object
            additionalProperties: true
          policy:
            type: object
            properties:
              false_positive_cost:
                type: number
                minimum: 0.01
                maximum: 1000
                default: 1
              false_negative_cost:
                type: number
                minimum: 0.01
                maximum: 1000
                default: 4
              block_threshold:
                type: number
                minimum: 0.01
                maximum: 0.99
          options:
            type: object
            properties:
              return_sanitized:
                type: boolean
                default: true
      Finding:
        type: object
        required:
        - category
        - severity
        - confidence
        - path
        - offset
        - evidence_sha256
        properties:
          category:
            type: string
          severity:
            enum:
            - medium
            - high
            - critical
          confidence:
            type: number
          path:
            type: string
          offset:
            type: integer
          evidence_sha256:
            type: string
      ShieldResponse:
        type: object
        properties:
          ok:
            const: true
          status:
            const: ANALYZED
          result:
            type: object
            properties:
              verdict:
                enum:
                - CLEAN
                - REVIEW
                - QUARANTINED
              allowed:
                type: boolean
              risk_score:
                type: number
              findings:
                type: array
                items:
                  $ref: '#/components/schemas/Finding'
              sanitized_payload: {}
              cleanliness_proof:
                type: object
                properties:
                  format:
                    const: hm-clean-v1
                  algorithm:
                    const: HMAC-SHA256
                  token:
                    type: string
      x-humanmirror-pricing-tier: granular_unit_0_001
x-humanmirror-canonical-bundle: https://humanmirror.fr/api/v1/agent/security-intelligence/