MCP Server for routing agent queries to optimal model tier based on complexity and SLA using a cascading router strategy
Single tool with basic schema and minimal descriptions. Tool name 'route_query' lacks action verb clarity (should be 'route_to_model' or similar). Input schema present but parameter descriptions are sparse (prompt, quality_sla, max_cost lack context on ranges, formats, and use cases). No output schema documented. No error handling guidance. No parameter validation rules stated. Complexity assessment logic is hardcoded keyword-matching, not exposed to agent for transparency.
Route agent query to optimal model tier based on complexity and SLA
Parameter descriptions missing or trivial. 'prompt' has no description; 'quality_sla' lacks range/format guidance (0.0-1.0 is in schema but not param description); 'max_cost' has no units or constraints.
No output schema documented. Agents cannot predict response structure (selected_tier, complexity_score, quality_sla, estimated_cost_per_million, expected_quality, expected_latency_ms). LLM must infer from example or fail.
Tool name 'route_query' is vague. Does not start with clear action verb. 'route_to_model', 'select_model_tier', or 'optimize_model_routing' would be clearer.
No error handling or recovery guidance. If complexity assessment fails or max_cost constraint is unsatisfiable, no error message tells agent what to do next.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 45 | 2026-07-28+ | v2 |
Tool description does not explain when to use it vs. alternatives, what prerequisites exist, or what the returned tier means operationally. LLM cannot reason about tool selection.