Project Aura has 14 tools with significant definition gaps. Tool naming is inconsistent (verb_noun pattern mostly followed, but some names are vague like 'get_phrase'). Descriptions exist but are often too brief (10-50 chars) and lack context for LLM selection. Input schemas are present with type definitions, but parameter descriptions are sparse or missing entirely. Output schemas are undocumented. Error handling is minimal, no recovery guidance or actionable error messages. The 'lookup_hint' and 'get_hint' tools have bilingual descriptions that are confusing. No tool declares permissions, audit trails, or idempotency guarantees. The green_tea_skill module tools lack clarity on what 'rating', 'intensity', and 'cooldown' mean to an LLM.

• Document output schemas for all tools. E.g., 'get_system_instructions' returns {type: string, length: number}; 'get_phrase' returns {phrase: string, rating: number, intensity: number}.
• Replace bilingual descriptions with English-only, LLM-optimized text (50-200 chars). E.g., 'lookup_hint: Given user message, husband state tags, and emotional temperature (0.0-1.0), return the best interaction strategy hint.'
• Add enum constraints to 'module_type' parameter: ['transcendence', 'comfort', 'admiration', 'vulnerability', 'ice_breaking', 'coquettishness']. Add range to 'min_intensity': 1-5.
• Implement actionable error messages. E.g., 'update_rating: phrase not found. Try increase_rating(keyword=...) to search by partial match.' Return {success: bool, message: string, updated_rating?: number}.
• Rename tools for clarity: 'get_phrase' → 'get_random_phrase'; 'lookup_hint' → 'lookup_strategy_hint'; 'get_hint' → 'get_strategy_phrase'. Add descriptions explaining when to call each.