How does a Hantera MCP tool get its input and output contract?
Hantera derives an MCP tool’s input schema from the target Reactor component’s parameter declarations and derives its output schema from the compiled result where the type can be represented.
The component is the source of truth for the tool’s supported parameters and result shape. If the contract cannot be represented safely as an MCP tool, the ingress is not shown in MCP discovery.
This reduces the risk that an AI-tool description drifts away from the code that actually runs.