MCP Proxy Service with Authentication, Logging, Billing and Load Balancing
ToolGate exposes only 2 tools with minimal schema documentation. listTools has a trivial description (5 words) and no input schema visible. callTool has a basic input schema but lacks output schema documentation, error handling guidance, and parameter descriptions are generic. Neither tool has descriptions meeting the 10-1024 character baseline (listTools is ~40 chars, callTool is ~80 chars but lacks depth). No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite callTool being a WRITE operation. No pagination support visible for listTools despite it returning potentially large tool lists. Error handling is absent, no recovery guidance, no categorization of retryable vs fatal errors.
Execute a tool from a connected MCP provider with specified arguments
List all available tools from connected MCP providers
callTool lacks output schema documentation. LLMs cannot plan downstream calls or extract returned tool results without knowing the response structure.
listTools description is 40 characters ('List all available tools from connected MCP providers'), below 10-1024 baseline. Lacks WHEN to use, WHAT it returns, and any prerequisites.
callTool input parameters 'name' and 'arguments' lack actionable descriptions. 'name' should specify format (e.g., 'provider_key/tool_name'), and 'arguments' should document expected structure and validation rules.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 38 | <=2025-11-25 | v2 |
callTool is marked WRITE but has no error handling guidance, no confirmation/dry-run pattern, and no tool annotations (destructiveHint). Agents cannot safely reason about side effects.