MCP Template
MCP Server

Tools

Available MCP tools and how to use them

Once connected, your AI assistant has access to these tools:

ToolDescription
greetGreet a user by name
config_showShow the full configuration
config_getGet a single config value
config_setSet a config override
doctorRun project health checks

greet

Greet a user by name, optionally shouting.

ParameterTypeDefaultDescription
namestring(required)Name to greet
shoutbooleanfalseUppercase the greeting
timesinteger1Repetition count

Example:

Input:

{ "name": "Alice", "shout": true, "times": 1 }

Result:

{ "message": "HELLO, ALICE!", "times": 1 }

config_show

Show the full project configuration. No parameters.

Example:

Input:

{}

Result:

{
  "config": {
    "llm_config": { "cache_enabled": true },
    "logging": { "verbose": false, "level": "INFO" }
  }
}

config_get

Get a single configuration value by dot-separated key.

ParameterTypeDescription
keystringDot-separated config path

Example:

Input:

{ "key": "llm_config.cache_enabled" }

Result:

{ "key": "llm_config.cache_enabled", "value": true }

config_set

Set a configuration override. The value is auto-coerced to the appropriate type.

ParameterTypeDescription
keystringDot-separated config path
valuestringNew value (auto-coerced to bool/int/float/string)

Example:

Input:

{ "key": "logging.verbose", "value": "true" }

Result:

{ "key": "logging.verbose", "coerced_value": true }

doctor

Run project health checks. Optionally auto-fix fixable issues.

ParameterTypeDefaultDescription
fixbooleanfalseAuto-fix fixable issues

Example:

Input:

{ "fix": false }

Result:

{
  "checks": [
    { "name": "Python version", "status": "pass", "message": "3.12.0" },
    { "name": "Deps synced", "status": "warn", "message": "dependencies out of sync", "fixable": true }
  ],
  "has_failures": false
}

Error Handling

If a tool call fails, the AI client receives an error message and can adjust:

{ "error": "Key not found: nonexistent.key" }

On this page