Loading ToolBench
MCP Server starter kit with ApiNow tools for searching and executing API endpoints on the Base blockchain network
Two tools with complete input schemas and descriptions, but significant gaps in parameter documentation and output schema clarity. apinow_search has 3 parameters with descriptions; apinow_execute has 6 parameters with descriptions. However, neither tool documents its output schema structure, limiting LLM ability to chain results. Parameter descriptions are present but lack constraint details (ranges, formats, validation rules). No error handling guidance provided to help LLMs recover from failures. Tool names follow verb_noun pattern (search, execute) which is good, but descriptions lack WHEN/WHY context for tool selection.
Executes a specific ApiNow endpoint by URL or by namespace/name, paying the required fee.
Searches ApiNow endpoints using natural language.
Output schemas not documented. Neither tool specifies what fields the LLM should expect in responses (e.g., apinow_search returns {success, data, txHash, txHashUrl} but this is not declared in tool definition). LLMs cannot plan downstream calls or extract chaining IDs without documented output structure.
apinow_execute has conditional required parameters (endpointUrl OR namespace+endpointName) but this mutual exclusivity is not formally declared in the schema. The description mentions it, but JSON Schema should use oneOf or explicit conditional logic. LLMs may pass both or neither.
Parameter descriptions lack constraint details. 'limit' and 'minScore' in apinow_search have no min/max bounds stated. 'method' enum is present but no default behavior documented. 'data' accepts additionalProperties:true with no guidance on structure. Unbounded parameters invite invalid LLM inputs.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 48 | 2026-07-28+ | v2 |
No error handling guidance. Handlers throw generic errors ('ApiNow search failed: ...') without categorizing as retryable, user-fixable, or fatal. LLMs receive no recovery hints (e.g., 'Check APINOW_WALLET_PKEY is set' or 'Retry with lower minScore').
Tool descriptions lack WHEN/WHY context. 'Searches ApiNow endpoints using natural language' does not explain when to call this vs apinow_execute, what prerequisites exist (wallet setup), or what the typical workflow is. Descriptions should guide tool selection.