Two tools with minimal descriptions (both under 150 chars) and vague naming. 'group-text-by-json' and 'text-to-json' lack action verbs and clarity about what they actually do. Schemas are present but parameters lack detail. Tool descriptions do not explain WHEN to use them, WHAT they return, or dependencies between them. No error handling guidance. No output schema documentation. Naming violates verb_noun convention, 'group-text-by-json' is a noun phrase, not an action.

• Rename tools to verb_noun format: 'extract_text_groups' (instead of 'group-text-by-json') and 'convert_text_to_json' (instead of 'text-to-json'). This signals action and helps LLMs disambiguate.
• Expand descriptions to 150 - 250 chars. Example: 'extract_text_groups: Analyzes unstructured text and groups it by JSON template keys. Returns a formatted text response ready for the convert_text_to_json tool. Use this first to prepare text for JSON conversion.'
• Document output schema in tool definitions. Both tools return {content: [{type: 'text', text: string}]}. Declare this so LLMs know what to expect.
• Add parameter constraints and examples. For 'template': 'Valid JSON object with string keys (e.g., {"brand": "", "price": ""}). Keys become grouping categories.' For 'text': 'Unstructured text output from extract_text_groups tool, formatted as key: value pairs.'
• Add error handling guidance in descriptions. E.g., 'If template is invalid JSON, returns error with hint to validate syntax. If text does not contain expected keys, returns partial results with available matches.'

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

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