naftiko: 1.0.0-alpha2
info:
  label: OpenAI Moderations — Moderations
  description: 'OpenAI Moderations — Moderations. Self-contained Naftiko capability for classifying text
    and images across harassment, hate, sexual, self-harm, violence, and illicit content categories.'
  tags:
  - Openai
  - Moderations
  - Content Safety
  created: '2026-05-22'
  modified: '2026-05-22'
binds:
- namespace: env
  keys:
    OPENAI_API_KEY: OPENAI_API_KEY
capability:
  consumes:
  - type: http
    namespace: moderations-moderations
    baseUri: https://api.openai.com/v1
    description: OpenAI Moderations business capability.
    resources:
    - name: moderations
      path: /moderations
      operations:
      - name: createmoderation
        method: POST
        description: Classify input text and/or images for policy categories. Supports omni-moderation-latest for multi-modal inputs and is free to use.
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
        inputParameters:
        - name: body
          in: body
          type: object
          required: true
    authentication:
      type: bearer
      token: '{{env.OPENAI_API_KEY}}'
  exposes:
  - type: rest
    namespace: moderations-moderations-rest
    port: 8080
    description: REST adapter for OpenAI Moderations.
    resources:
    - path: /v1/moderations
      name: moderations
      operations:
      - method: POST
        name: createmoderation
        description: Create a moderation classification.
        call: moderations-moderations.createmoderation
        with:
          body: rest.body
        outputParameters:
        - type: object
          mapping: $.
  - type: mcp
    namespace: moderations-moderations-mcp
    port: 9090
    transport: http
    description: MCP adapter for OpenAI Moderations.
    tools:
    - name: openai-create-moderation
      description: Classify text and/or image inputs against OpenAI content policy categories.
      hints:
        readOnly: true
        destructive: false
        idempotent: true
      call: moderations-moderations.createmoderation
      with:
        body: tools.body
      outputParameters:
      - type: object
        mapping: $.