24 tools with consistent naming (verb_noun pattern: kreta_*, classroom_*) and Hungarian descriptions. All tools have input schemas with type definitions and parameter descriptions. However, descriptions are brief (avg ~60 chars, well below the 194-char baseline), lack WHEN-to-use guidance, and output schemas are not documented. Error handling is present but minimal. Tool composition is sound (single responsibility), but parameter descriptions could be more actionable with format/constraint details.

• Expand tool descriptions to 100 - 200 chars. Include WHAT the tool does, WHEN to use it (vs similar tools), and any prerequisites. E.g., 'kreta_evaluations: Fetch student grades and marks. Use this to answer questions about academic performance. Requires a connected child profile.'
• Document output schemas for all tools. Specify which fields are returned, their types, and which are safe to pass to downstream tools. E.g., kreta_class_groups should document that it returns group_id, group_name, and teacher_name so classroom_coursework can accept group_id.
• Add explicit min/max constraints to numeric parameters. E.g., 'limit: number (1 - 100, default 20)' instead of 'Legfeljebb ennyi elem.' This prevents LLMs from passing absurd values.
• Add recovery guidance to error-prone tools. E.g., kreta_add_child: 'If the school is not found, call kreta_list_children to see available schools.' This guides LLMs toward self-correction.
• Clarify date range semantics. E.g., 'start_date and end_date define an inclusive range. If end_date < start_date, the tool returns an error. Dates must be within the current school year.'