Mock Component¶

The Mock component executes tools with configurable responses. Load any tool schema and define mock responses, fixtures, or let the system generate default outputs.

Use Cases¶

Test agent workflows without real API access
Create training data for external services (GitHub, Slack, databases)
Define deterministic responses for reproducible datasets

Loading Tool Schemas¶

From JSON/YAML¶

curl -X POST http://localhost:3000/mock/load-schema \
  -H "Content-Type: application/json" \
  -d '[
    {
      "name": "get_weather",
      "description": "Get weather for a location",
      "inputSchema": {
        "type": "object",
        "properties": {
          "location": {"type": "string"}
        },
        "required": ["location"]
      },
      "mock_response": {"temperature": 72, "condition": "sunny"}
    }
  ]'

From MCP Server¶

Pull tools from a Model Context Protocol server:

curl -X POST http://localhost:3000/mock/pull \
  -H "Content-Type: application/json" \
  -d '{"url": "http://localhost:8000/mcp"}'

Configuration¶

generation:
  tools:
    spin_endpoint: "http://localhost:3000"
    tools_endpoint: "http://localhost:3000/mock/list-tools"
    tool_execute_path: "/mock/execute"

Use tools_endpoint to load tools and tool_execute_path for execution.

Mock Responses¶

Default Behavior¶

Without a mock_response, the component echoes the tool call:

{
  "tool": "get_weather",
  "arguments": {"location": "Seattle"},
  "mock": true
}

Custom Response¶

Define a static response in the schema:

{
  "name": "get_weather",
  "mock_response": {
    "temperature": 72,
    "condition": "sunny"
  }
}

Template Interpolation¶

Use {{argument_name}} to include call arguments in responses:

{
  "name": "get_weather",
  "mock_response": {
    "location": "{{location}}",
    "temperature": 72
  }
}

Calling get_weather(location="Seattle") returns:

{"location": "Seattle", "temperature": 72}

Fixtures¶

Fixtures return specific responses based on argument matching:

curl -X POST http://localhost:3000/mock/add-fixture \
  -H "Content-Type: application/json" \
  -d '{
    "name": "get_weather",
    "match": {"location": "Seattle"},
    "response": {"temperature": 55, "condition": "rainy"}
  }'

Now get_weather(location="Seattle") returns the rainy fixture, while other locations use the default mock response.

Multiple Fixtures¶

Add fixtures for different argument combinations:

# Seattle fixture
curl -X POST .../add-fixture -d '{
  "name": "search_code",
  "match": {"repo": "main", "query": "TODO"},
  "response": {"matches": ["file1.py:10", "file2.py:25"]}
}'

# Different repo
curl -X POST .../add-fixture -d '{
  "name": "search_code",
  "match": {"repo": "testing"},
  "response": {"matches": []}
}'

API Reference¶

Endpoint	Method	Description
`/mock/load-schema`	POST	Load tool definitions
`/mock/pull`	POST	Pull from MCP server
`/mock/execute`	POST	Execute a tool
`/mock/update-response`	POST	Update tool's mock response
`/mock/add-fixture`	POST	Add argument-specific fixture
`/mock/list-tools`	GET	List loaded tools
`/mock/clear`	POST	Clear all tools

Execute Request¶

{
  "name": "tool_name",
  "arguments": {"arg1": "value1"}
}

Note: Mock execute uses name and arguments, unlike VFS which uses tool and args.

Example: GitHub Tools¶

# Load GitHub tool schemas
generation:
  tools:
    spin_endpoint: "http://localhost:3000"
    tools_endpoint: "http://localhost:3000/mock/list-tools"
    tool_execute_path: "/mock/execute"

Then load schemas and fixtures:

# Load GitHub tools
curl -X POST http://localhost:3000/mock/load-schema \
  -d '[{"name": "create_issue", "description": "Create GitHub issue", ...}]'

# Add fixture for specific repo
curl -X POST http://localhost:3000/mock/add-fixture \
  -d '{
    "name": "create_issue",
    "match": {"repo": "myorg/myrepo"},
    "response": {"issue_number": 42, "url": "https://github.com/myorg/myrepo/issues/42"}
  }'