keou-mcp has 9 tools with basic descriptions and parameter schemas, but significant gaps prevent production use. Tool names lack clear verb-noun structure (keou_generate_image vs generate_image_with_keou). Descriptions are present but generic (avg ~80 chars, below 194-char baseline). Parameter descriptions exist but lack constraints, ranges, and format guidance. No output schemas documented. Error handling is minimal, no recovery guidance or actionable error messages. No tool annotations (readOnlyHint/destructiveHint). STDIO transport caps protocol readiness at 50.

• Document output schemas for all tools. E.g., keou_generate_image should return {taskId: string, status: string, imageUrl?: string, estimatedTime?: number}. This enables LLMs to chain tools and extract results.
• Add min/max constraints and format guidance to numeric/string parameters. E.g., 'scale: number (min 2, max 8, e.g., 4 for 4x upscaling)', 'width: number (pixels, 256 - 2048, default 512)'.
• Implement error recovery guidance. E.g., 'If API key is missing, call keou_status_keys() to check which providers are configured. Sign up at https://kie.ai if needed.'
• Rename tools to follow verb_noun convention without brand prefix. E.g., 'generate_image', 'generate_video', 'upscale_image', 'get_generation_status', 'check_api_keys', 'generate_product_pack', 'apply_brand_styling', 'generate_voice', 'generate_sound_effect'.
• Add tool annotations via MCP SDK. Mark destructive tools (generate_*, pack, brand_kit) with destructiveHint=true; mark read-only tools (get_status, status_keys) with readOnlyHint=true; mark idempotent tools with idempotentHint=true.