Three tools with Ukrainian descriptions and basic schemas. All tools lack English descriptions (critical for LLM selection in English-speaking contexts). Schemas present but minimal, parameters have types but descriptions are sparse or missing. ask_colleague and ask_colleagues have timeout parameters but no guidance on when to use 300-600s vs 120s default. ask_manager lacks any parameter description. No error handling guidance, no output schema documentation, no idempotency markers. Tool names use verb_noun pattern (ask_*) which is good, but descriptions are in Ukrainian only, limiting accessibility.

• Add English descriptions (50-200 chars) for all three tools explaining WHAT they do, WHEN to use them, and WHAT they return. Example: 'ask_colleague: Delegate a task to a named bot. Use when you need specialized expertise or parallel processing. Returns the bot's response and maintains session context across calls.'
• Document output schemas for each tool. Specify: response structure, whether it includes bot_name/status/error fields, and what fields downstream tools can rely on.
• Add parameter descriptions for ask_manager's 'question' field: expected format, length constraints, and examples of valid questions.
• Replace vague timeout guidance with explicit rules: 'Use 120s (default) for simple queries. Use 300s for multi-step tasks. Use 600s only if the bot will spawn sub-agents.' Add min/max constraints (e.g., 10-3600).
• Add error handling guidance: 'If bot is unavailable, try ask_colleagues with a fallback bot. If timeout occurs, increase timeout and retry. If bot returns an error, ask_manager for escalation.'