Two tools with complete schemas and bilingual descriptions. Naming follows verb_noun pattern (list_, get_). Descriptions are adequate (100-150 chars) but lack actionable context for LLM selection. Schemas are well-formed with enums and required fields. No output schema documentation. No error handling guidance. No parameter descriptions for 'lang' beyond enum values. Tool composition is sound (list → get chain), but descriptions don't explain when to use each tool or what downstream actions follow.

• Document output schemas for both tools. E.g., list_thinking_frameworks returns {frameworks: [{id, name, name_en, when, when_en, groupName, groupName_en}]}. get_thinking_framework returns {name, name_en, groupName, groupName_en, when, when_en, body, body_en, note, note_en, rules, rules_en}.
• Rewrite tool descriptions to answer: What does it do? When should the LLM call it? What does it return? Example: 'list_thinking_frameworks: Discover which of 12 thinking frameworks fits the user's problem (ambiguous question, concept learning, stuck on a hard problem, decision-making, career/life planning). Returns framework IDs, names, and applicability hints. Call this first, then use get_thinking_framework to fetch the full prompt and execution rules.'
• Add per-parameter descriptions. For 'id' in get_thinking_framework, add: 'Framework identifier. Must be one of the 12 IDs returned by list_thinking_frameworks (e.g., socratic, first-principles, steelman).'
• Implement structured error responses. When 'id' is invalid, return JSON: {error: 'Unknown framework ID', available_ids: [...], suggestion: 'Call list_thinking_frameworks to see all available frameworks.'}
• Separate bilingual text in descriptions. Use English descriptions for LLM clarity, and handle language selection via the 'lang' parameter only. Current inline mixing (zh/en) reduces readability.