Quran MCP has well-structured tool definitions with clear naming (verb_noun pattern) and comprehensive descriptions (avg 180 chars). All 4 tools have proper input schemas with typed parameters and descriptions. However, output schemas are not formally documented, responses are inferred from code. Parameter descriptions lack format constraints (e.g., edition enum values, surah range validation). Error handling is minimal; API errors return generic {error, message} without recovery guidance. No tool annotations (readOnlyHint, idempotentHint) despite all tools being read-only. Missing pagination support for search_quran despite potentially large result sets.

• Add formal outputSchema to each tool definition documenting return structure (e.g., list_surahs returns array of {number, name, english_name, english_translation, ayahs, revelation_type}).
• Replace edition parameter description with enum constraint: edition: {type: 'string', enum: ['en.asad', 'en.sahih', 'quran-uthmani', ...], description: 'Edition identifier...'}.
• Add pagination to search_quran: include limit (default 20, max 100) and offset parameters; return {count, total, matches, next_offset}.
• Enhance error handling: return {error: 'invalid_edition', message: 'Edition "xyz" not found. Valid editions: en.asad, en.sahih, quran-uthmani', retryable: false} to guide LLM recovery.
• Add tool annotations in inputSchema: readOnlyHint: true for all tools to signal safety and retry eligibility to agents.