Loading ToolBench
MCP server for Minter.io social media analytics platform, providing access to Instagram, TikTok, X, Facebook, LinkedIn, and Threads data
Critical failures across all three tools. No input schemas are visible in the source code, only tool names and descriptions are provided. The server exposes tools via a remote HTTP endpoint (https://minter.io/mcp) without publishing schema definitions in the repository. Tool descriptions are present but generic (10-60 chars), lacking actionable context for LLM selection. No parameter documentation, no output schema specification, and no error handling guidance. This is a remote MCP server where the actual tool definitions are hosted externally and not inspectable from the repo.
List every report tracked on Minter.io
Query social media analytics data for specified accounts and date ranges
Discover available fields and metrics for a given report
No input schemas visible. Tool definitions are hosted remotely at https://minter.io/mcp and not inspectable from the repository. Cannot verify parameter types, constraints, or required fields.
Tool descriptions are too short (10-60 chars) and lack actionable context. 'List every report tracked on Minter.io' does not explain when to call this vs field_discovery, what structure is returned, or how to use results downstream.
No output schema documentation. LLMs cannot plan downstream tool calls without knowing what fields accounts_discovery returns (e.g., does it return account_id, account_name, platform, metrics_available?).
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 38 | 2026-07-28+ | v2 |
Naming lacks clarity on parameter expectations. 'data_query' is vague, does it accept account_id or account_name? date_start or start_date? No parameter documentation visible.
No error handling guidance. If a user requests data for a non-existent account or invalid date range, the tool provides no recovery hints (e.g., 'Call accounts_discovery first to list valid accounts').