Loading ToolBench
AI Study Helper Agent for generating summaries and questions from study topics
This MCP server has critical definition quality gaps. While it implements 3 tools with basic FastAPI routing, the tools lack proper MCP schema registration, detailed parameter descriptions, and formal output schema documentation. Tool names lack action verbs (naming: summary, questions, answers instead of generate_summary, generate_questions, provide_answers). Parameter descriptions are minimal or absent. Output schemas are not formally documented. The codebase shows tool implementations but no evidence of proper MCP Tool registration with complete JSON Schema. The tools operate as REST endpoints rather than MCP-compliant tool definitions with structured input/output contracts.
Provides answers for questions
Generates study questions
Generates summary and key points
Tool names lack action verbs. 'summary', 'questions', 'answers' are nouns, not verb_noun patterns like 'generate_summary', 'list_questions', 'provide_answers'. LLMs rely on verb_noun structure to infer intent from the name alone.
Parameter descriptions are minimal or missing. The 'topic' parameter in summary and questions tools has a description, but answers tool's 'questions' array parameter lacks any explanation of what format is expected, what constraints apply, or how the tool uses it.
No formal output schema documentation. Tool implementations return Python dicts (summary + key_points, questions list, answers list) but there is no visible JSON Schema or formal documentation of return types, field names, or data types. This forces LLMs to guess what fields will be available.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 27 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 15 | - | v1 |
No MCP tool registration visible in source. The FastAPI router in app/api/study.py handles HTTP endpoints but does not show formal MCP Tool definitions with proper schema registration. Tools appear to be inferred from class definitions rather than explicitly registered with MCP.
Tool descriptions are generic. 'Generates summary and key points', 'Generates study questions', 'Provides answers for questions' lack context for when to use each tool vs similar alternatives. No mention of prerequisites, expected input format, or when the LLM should select this tool over others.
No error handling guidance. Tools may fail if LLM is not initialized, API keys are missing, or input is invalid, but there is no error response schema or recovery guidance for the agent.
answers tool accepts array of strings but does not specify expected array size, string length limits, or what constitutes a valid question. LLMs may pass malformed or excessively long arrays.