Server has 9 tools with basic definitions but significant quality gaps. Tool names lack action verbs (Add, Sub, CheckAngel, QueryAngelType, QueryGPUSpecs, GenerateText, CreateEmbedding, LoadModel, UnloadModel). Descriptions are present but many are vague or non-English (Chinese descriptions for GPU/model tools). Parameter descriptions exist but are minimal. Input schemas are visible in macro definitions but lack formal JSON Schema with type constraints. No output schemas documented. Error handling is minimal, tools return text responses without structured error guidance. No tool annotations (readOnlyHint/destructiveHint). Three tools (GenerateText, CreateEmbedding, LoadModel, UnloadModel) marked WRITE but lack confirmation/dry-run patterns. CheckAngel and QueryAngelType are low-value demo tools that should be removed.

• Rename tools to verb_noun format: 'add_numbers', 'subtract_numbers', 'get_gpu_specs', 'generate_text_with_model', 'create_text_embedding', 'load_model', 'unload_model'.
• Document output schemas for all tools. E.g., Add returns {result: number, operation: 'addition'}. GenerateText returns {text: string, model: string, tokens_used: number}.
• Add optional 'dry_run' parameter to GenerateText, CreateEmbedding, LoadModel, UnloadModel. When true, return what would happen without executing.
• Translate all descriptions to English. Chinese descriptions block LLM comprehension in English-first agents.
• Remove CheckAngel and QueryAngelType or move to a separate demo/test server. They add no production value and clutter the tool namespace.

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

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