Server has 4 tools with complete input schemas and descriptions. Naming follows verb_noun convention (generate_image, edit_image, get_image_result). Descriptions are present and moderately detailed (80-180 chars), explaining what each tool does and its return value. However, parameter descriptions are sparse or missing for several fields (e.g., 'watermark' lacks explanation of what the watermark does; 'response_format' description is minimal). Output schemas are not documented, LLMs cannot predict the structure of results. Error handling exists but is minimal (generic error messages without recovery guidance). No tool annotations (readOnlyHint/destructiveHint) despite clear write/read semantics. Job store pattern is reasonable but not idempotent-safe for retries.

• Add explicit output schema documentation to each tool. E.g., generate_image returns {request_id: string}; get_image_result returns {state: 'pending'|'succeeded'|'error', data?: Array<{url: string, size?: string}>, error?: string}.
• Expand parameter descriptions: explain what 'watermark' does (adds AI-generated watermark to image), clarify 'response_format' (url returns 24h-expiry URLs; b64_json returns base64-encoded PNG), document model quality/speed tradeoffs.
• Add tool annotations: mark generate_image, edit_image, generate_image_group with destructiveHint=true; mark get_image_result with readOnlyHint=true.
• Enhance error messages with recovery guidance. E.g., 'Job not found (expired after 10 minutes). Resubmit with generate_image() to start a new request.' or 'API rate limit exceeded. Retry in 30 seconds.'
• Document the 10-minute job TTL in get_image_result description and explain that expired jobs cannot be recovered.