add get_variables tool for viewing workspace configuration

[zzstoatzz]

overview

add read-only tool to query variables (workspace-level key-value configuration)

what are variables?

variables are workspace-level mutable json storage:

• use cases: configuration values, feature flags, environment-specific settings
• key features: simple name-based access, optional tags, any json-serializable value
• scope: shared across all flows in workspace

proposed tool

@mcp.tool
async def get_variables(
    name: str | None = None,
    limit: int = 50,
) -> VariablesResult:
    """get variables (workspace-level configuration).
    
    variables are mutable json values shared across flows, often used
    for configuration, feature flags, or environment-specific settings.
    
    examples:
        - list all: get_variables()
        - get specific: get_variables(name="api_url")
    """

implementation notes

types to add (types.py):

class VariableDetail(TypedDict):
    id: str
    name: str
    value: Any  # json value
    tags: list[str]
    created: str | None
    updated: str | None

class VariablesResult(TypedDict):
    success: bool
    count: int
    variables: list[VariableDetail]
    error: str | None

client code (_prefect_client/variables.py):

async def get_variables(name: str | None = None, limit: int = 50) -> VariablesResult:
    async with get_client() as client:
        if name:
            variable = await client.read_variable_by_name(name)
            variables = [variable] if variable else []
        else:
            variables = await client.read_variables(limit=limit)
        
        return {
            "success": True,
            "count": len(variables),
            "variables": [...],
            "error": None,
        }

value

enhances observability into workspace configuration - useful for checking current settings, feature flags, and environment-specific values

note on mutations

keeping this read-only consistent with server design. users can mutate via cli/ui when needed.