neo-mcp defines 4 tools with complete input schemas and descriptions. However, descriptions are minimal (10-50 chars) and lack actionable context for LLM selection. Tool names lack action verbs (txt2img, img2img, anlas, bgkill are noun-heavy or abbreviated). No output schemas documented. Parameters have types and descriptions, but lack constraints (enums, ranges, format hints). Error handling and recovery guidance absent. Composition is reasonable (each tool has one responsibility), but descriptions do not explain WHEN to use each tool or what to expect in responses.

• Rename tools with action verbs: 'generate_image_from_text' (txt2img), 'generate_image_from_image' (img2img), 'remove_background' (bgkill), 'check_anlas_balance' (anlas). This helps LLMs disambiguate intent.
• Expand descriptions to 80-150 chars with WHEN/WHY context. Example: 'Generate images from text prompts using Stable Diffusion. Use this for creative generation; use img2img for refinement or style transfer. Returns image URL and metadata.'
• Document output schemas in tool descriptions or a separate schema section. Example: 'Returns {image_url: string, seed: integer, generation_time_ms: integer}'.
• Add enums for constrained parameters. Example: output_format: enum [png, jpeg, jxl, webp]; sampler_name: enum [euler, dpm++, ...]; model_name: enum [General, Portrait, Anime-Lite].
• Add error handling guidance. Example: 'If generation times out, retry with fewer steps or lower resolution. If VRAM error, reduce batch size or use a smaller model.'

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry