MCP Server for converting JoinQuant strategies to Ptrade format
Single tool with moderate structural issues. Tool name is action-verb-based (convert_), which is correct, but description quality is mixed. Parameter descriptions exist but lack detail on formats and constraints. Input schema present but underdocumented. No output schema visible in source. Error handling approach not evident from source. Overall follows basic patterns but has significant gaps typical of community servers (median ~50).
将聚宽(JoinQuant)策略代码转换为Ptrade格式。可以直接输入代码字符串,或提供文件路径(支持.py/.txt等格式)。自动识别并转换所有API调用,生成Ptrade可用的.py文件和详细转换报告。
Missing output schema documentation. Tool description does not specify what fields are returned (convertedCode, changes, warnings, success visible in code but not documented for LLM consumption).
Parameter descriptions lack format constraints and dependencies. 'code' and 'filepath' are mutually exclusive (noted as 'second-choice' in Chinese) but not explicitly stated in English parameter descriptions. No description of expected output_dir behavior when omitted.
No required parameters declared in schema (required array is empty). LLM unclear whether code/filepath are optional or whether one must be provided. Input validation rules not visible in source.
Error handling and recovery guidance absent from tool description. No guidance on what happens if conversion fails, what errors are possible, or how to respond.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 46 | <=2025-11-25 | v2 |
| 2026-03-09 | F | 30 | 2024-11-05+ | v1 |
Tool description mixes English and Chinese, reducing clarity for English-speaking LLM agents. International tool descriptions should be consistently in a single language.