Two tools with basic schemas and descriptions, but significant gaps in parameter documentation, output schema clarity, and error handling. Tool names lack action verbs (check_channel, compare_channels are descriptive but not verb-first). Descriptions are present but generic. Parameters have types but lack detailed constraints and validation guidance. No documented output schemas. Error handling is minimal, no recovery guidance or actionable error messages. The checker.py implementation shows basic logic but no structured error responses or input validation.

• Rename tools to verb-first: 'analyze_channel' or 'evaluate_channel' instead of 'check_channel'; 'rank_channels' or 'compare_channels_by_reach' instead of 'compare_channels'. This helps LLMs parse intent from the name alone.
• Document output schemas in tool descriptions. E.g., 'Returns: {channel: str, median_views: int, reach_ratio: float, cpm: float, verdict: str (🟢/🟡/🔴), reasons: list[str]}'.
• Add parameter constraints to descriptions. E.g., 'price: cost per post in rubles (must be > 0); max_cpm: CPM threshold in rubles (default 500, typical range 100-1000); channel: Telegram channel URL (https://t.me/...) or handle (@channelname)'.
• Implement structured error responses with recovery guidance. E.g., {'error': 'Channel has no public posts', 'recovery': 'Try a different channel or verify it is public', 'retryable': false}.
• Add input validation: sanitize channel parameter against URL/handle patterns, validate price and max_cpm are positive numbers, and return clear error messages for invalid inputs.