Loading ToolBench
MCP Server for Zerocracy - provides management advice and product development insights
The server defines a single tool 'give_management_advice' with significant quality gaps. The tool name lacks a clear action verb (should be 'get_management_advice' or similar), the input schema uses Zod objects directly rather than proper JSON Schema with type annotations, parameter descriptions are missing entirely, and there is no documented output schema. The tool description, while detailed (245 chars), lacks clarity on what the LLM should expect in return and when specifically to invoke it. The schema registration uses a Zod object ({ concern: z.string(), product: z.string() }) without explicit type/description metadata visible in the registration call, making parameter intent opaque to the LLM. No error handling documentation, no pagination, no structured output schema.
Analyze the most critical concern related to product development, providing clear insights into the underlying issues, identifying areas for improvement, and outlining specific corrective and preventive actions. The objective is to directly address and resolve this concern, helping the team become more productive, focused, and result-oriented. These recommendations can later be converted into new GitHub issues or added as comments to existing issues or pull requests.
Tool name lacks action verb, 'give_management_advice' is ambiguous. LLMs cannot infer what action triggers. Should be 'get_management_advice', 'request_management_advice', or 'analyze_management_concern'.
Parameters lack descriptions in tool registration. The 'concern' and 'product' parameters are undocumented, LLMs cannot determine expected format, length constraints, or example values.
No output schema documented. The tool returns text content but LLMs have no metadata about the structure, format, or expected fields. Breaks tool chaining and downstream reasoning.
Input schema uses Zod objects without explicit JSON Schema type/description metadata in the registration. Zod v4 compatibility requires @ts-ignore and type depth workarounds, indicating the schema may not be properly serialized for LLM consumption.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-21 | F | 39 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 41 | - | v1 |
No error handling guidance. The tool calls an external API (baza) that can fail with network errors, HTTP errors, or missing ZEROCRACY_TOKEN. No recovery instructions or error categorization for the LLM.
Secrets in implementation. The ZEROCRACY_TOKEN is read from environment and passed to external API, but the baza function throws a raw error if missing. Should validate at server startup and fail gracefully.
Tool description does not state what the tool returns or what downstream actions are possible with the result. LLMs need to know: Is this a list of recommendations? A single action plan? Structured data or free text?