Three tools with complete input schemas and descriptions. Naming follows verb_noun convention (get_*, search_*). Descriptions are adequate (80-150 chars) but lack actionable context for LLM selection. All parameters have types and descriptions. No output schemas documented. Error handling not visible in provided code. Missing guidance on when to use each tool vs. alternatives, and no recovery hints for common failures.

• Document output schemas for all three tools. Example: 'Returns {snapshot: {feeds: {[feedName]: {items: [{id, timestamp, title, summary}], count}}, theater, timestamp}}'. This enables LLMs to chain calls and extract IDs.
• Expand descriptions to 150-250 chars with selection context. Example for get_snapshot: 'Fetch a consolidated view of all feeds for a theater. Use this for broad situational awareness. Use search_intel instead if you need to find specific events by keyword.'
• Add enum constraint to get_snapshot 'include' parameter items. Ensure BIG_BOSS_MCP_FEEDS values are explicitly listed in the schema, not just in code comments.
• Document error cases and recovery paths in descriptions. Example: 'If theater is invalid, available options are: middle-east, ukraine. If a feed name is unrecognized, call get_snapshot without the include parameter to see available feeds.'
• Add default value documentation for optional parameters. Example: 'limit defaults to 20 if omitted; maximum 50 to prevent context window exhaustion.'