RUNaDIVA

The two tools have clearly distinct targets: one evaluates a caller-supplied economic decision, while the other evaluates a single fixed synthetic sandbox case. Their descriptions make the boundary explicit, leaving no realistic ambiguity for an agent.

Both tool names use the same verb_noun pattern with snake_case ('evaluate_economic_decision', 'evaluate_sandbox_case'). The convention is consistent across the entire set.

Two tools is on the thin side for a dedicated server, though it may be appropriate if the sole purpose is decision evaluation. The set feels minimal and could benefit from additional operations like retrieving or listing sandbox cases.

For a stateless economic decision evaluation service, the surface covers the core action of evaluating a decision and provides a sandbox counterpart. Minor gaps exist around managing or discovering sandbox cases, but the primary workflow is complete.