Naftiko · Capability

Aiml Data Platform Context Composition

A -Global-Tech-shaped capability across AI/ML + data platform APIs implementing Compose AI Context.

Run with Naftiko NaftikoAI/MLData PlatformComposition

What You Can Do

POST
Compose ai context — Pull data-platform features (Snowflake) plus model metadata (MLflow) into one shaped AI context object.
/ai-context

MCP Tools

compose-ai-context

Compose AI context from data-platform + ML registry.

get-registered-model

Get a registered MLflow model.

read-only

Capability Spec

aiml-data-platform-context-composition.yaml Raw ↑
naftiko: 1.0.0-alpha2
info:
  title: Aiml Data Platform Context Composition
  description: A -Global-Tech-shaped capability across AI/ML + data platform APIs implementing Compose AI Context.
  tags: [Naftiko, AI/ML, Data Platform, Composition]
  created: '2026-05-01'
  modified: '2026-05-04'
binds:
- namespace: snowflake-env
  description: Snowflake account, user, and PAT.
  keys: {SNOWFLAKE_ACCOUNT: SNOWFLAKE_ACCOUNT, SNOWFLAKE_TOKEN: SNOWFLAKE_TOKEN}
- namespace: mlflow-env
  description: MLflow tracking-server bearer token.
  keys: {MLFLOW_TOKEN: MLFLOW_TOKEN, MLFLOW_HOST: MLFLOW_HOST}
capability:
  consumes:
  - namespace: snowflake
    type: http
    baseUri: https://{{SNOWFLAKE_ACCOUNT}}.snowflakecomputing.com
    authentication: {type: bearer, token: '{{SNOWFLAKE_TOKEN}}'}
    resources:
    - name: query
      path: /api/v2/statements
      operations:
      - {name: run-sql, method: POST, description: Run a SQL statement.}
  - namespace: mlflow
    type: http
    baseUri: https://{{MLFLOW_HOST}}
    authentication: {type: bearer, token: '{{MLFLOW_TOKEN}}'}
    resources:
    - name: registered-models
      path: /api/2.0/mlflow/registered-models/get
      operations:
      - name: get-registered-model
        method: GET
        inputParameters: [{name: name, in: query}]
    - name: experiments
      path: /api/2.0/mlflow/experiments/search
      operations:
      - {name: search-experiments, method: POST}
  exposes:
  - type: rest
    address: 0.0.0.0
    port: 8080
    namespace: aiml-data-platform-context-composition-rest
    description: REST surface composing data-platform query + ML model context into one call.
    resources:
    - name: ai-context
      path: /ai-context
      operations:
      - method: POST
        name: compose-ai-context
        description: Pull data-platform features (Snowflake) plus model metadata (MLflow) into one shaped AI context object.
        call: snowflake.run-sql
  - type: mcp
    address: 0.0.0.0
    port: 3010
    namespace: aiml-data-platform-context-composition-mcp
    description: MCP server exposing composed AI context retrieval.
    tools:
    - {name: compose-ai-context, description: Compose AI context from data-platform + ML registry., call: snowflake.run-sql}
    - name: get-registered-model
      description: Get a registered MLflow model.
      hints: {readOnly: true}
      inputParameters: [{name: name, type: string, required: true}]
      call: mlflow.get-registered-model
  - type: skill
    address: 0.0.0.0
    port: 3011
    namespace: aiml-data-platform-context-composition-skills
    description: Skill bundle for AI context composition.
    skills:
    - name: aiml-data-platform-context-composition
      description: Compose AI context from data + model platforms.
      location: file:///opt/naftiko/skills/aiml-data-platform-context-composition
      allowed-tools: compose-ai-context,get-registered-model
      tools:
      - {name: compose-ai-context, from: {sourceNamespace: aiml-data-platform-context-composition-mcp, action: compose-ai-context}}
      - {name: get-registered-model, from: {sourceNamespace: aiml-data-platform-context-composition-mcp, action: get-registered-model}}