Two tools with minimal schema and weak descriptions. generate_test_case has a single boolean parameter with description present but vague guidance embedded in docstring rather than parameter annotation. markdown_to_xmind_json accepts no parameters and lacks actionable description. Both tools return unstructured strings (prompt templates) rather than structured output. No error handling, no output schema documentation, no pagination or result limits. Tool names lack clarity on what 'generate' produces (markdown? file? prompt?). Descriptions are in Chinese, reducing LLM comprehension in English-first agents.

• Translate all tool descriptions and parameter annotations to English. Use clear, concise descriptions (50 - 200 chars) explaining WHAT the tool does, WHEN to call it, and WHAT it returns.
• Document output schema explicitly. E.g., generate_test_case returns {"markdown_content": string, "xmind_file_path": string | null}. markdown_to_xmind_json returns {"conversion_rules": string}.
• Rename tools for clarity: 'generate_test_case' → 'generate_test_case_markdown' or 'create_test_case_prompt'. 'markdown_to_xmind_json' → 'get_markdown_to_xmind_conversion_rules'.
• Add error handling: catch file I/O errors and return structured error with recovery hint, e.g., {"error": "Prompt template not found", "recovery": "Ensure prompt files are installed in package directory"}.
• Add parameter constraints: make is_need_xmind an enum or add explicit description: 'Set to true only if user explicitly requests XMind/mind map format; otherwise false.'

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry