Loading ToolBench
AI-powered study companion with MCP and Gemini integration for Indian students (Grades 5-10), supporting CBSE, ICSE, and IGCSE curricula
StudyBuddy MCP presents a STDIO-only educational tool suite with significant definition quality gaps. While all 5 tools have descriptions and visible schemas, the descriptions are generic and lack LLM-optimized guidance. Parameters are well-typed with enums where appropriate (board, grade ranges), but descriptions are minimal (10-40 chars for many params). Error handling is absent from visible code. Tool naming follows verb_noun pattern (e.g., 'studybuddy_explain_topic') but the 'studybuddy_' prefix is redundant given the server context. No output schemas are documented. Tool 5 (studybuddy_quiz_me) is marked WRITE but lacks any confirmation/dry-run mechanism. The codebase shows working Gemini integration in trainbot_tools.py but studybuddy server.py implementation is not fully visible, making schema inference incomplete.
Turn boring topics into fun, engaging stories with relatable characters. Makes learning memorable and enjoyable.
Explain any topic at grade-appropriate level aligned with CBSE/ICSE/IGCSE curriculum. Returns structured explanation with key points and examples.
Generate practice questions for any topic with varying difficulty levels and question types (MCQ, short answer, numerical).
Generate a 10-question quiz on any topic. Tracks previously asked questions to avoid repeats. Returns questions with answers and explanations for self-assessment.
Solve math/science problems with detailed step-by-step explanations, showing all work and reasoning.
Parameter descriptions are minimal (10-40 chars). 'The topic to explain', 'Subject name', 'Grade level' lack context for when/how to use them or format constraints.
Output schemas are not documented. LLMs cannot infer what fields/structure to expect from responses, blocking downstream chaining and result extraction.
Tool 'studybuddy_quiz_me' is marked WRITE (destructive) but has no confirmation mechanism, dry-run option, or recovery guidance. Agents could inadvertently generate duplicate quiz sessions.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 45 | <=2025-11-25 | v2 |
| 2026-03-09 | F | 42 | - | v1 |
Tool names redundantly prefix 'studybuddy_' (e.g., 'studybuddy_explain_topic'). Within the StudyBuddy MCP context, shorter names like 'explain_topic' are clearer. The prefix adds 11+ characters without semantic value.
No error handling strategy is visible. Tools lack guidance on retryability, user-fixable vs. fatal errors, or recovery paths. E.g., what if Gemini API rate-limits or the topic is invalid?
studybuddy_solve_step_by_step lacks a 'board' parameter present in other tools. This inconsistency may confuse LLMs about curriculum alignment and limits the tool's utility for multi-board contexts.