Opinion MCP has 9 well-named, read-only tools with consistent verb_noun naming (GET_*). All tools have descriptions (avg ~80 chars) and input schemas with typed parameters. However, descriptions are terse and lack context for LLM selection. Output schemas are not documented, LLMs cannot plan downstream calls or extract chaining IDs. No error handling guidance, no pagination documentation despite tools accepting limit/page params. Parameter descriptions are minimal (e.g., 'The token ID to get the latest price for' lacks format/constraint details). Missing idempotency hints and recovery guidance.

• Document output schemas for all 9 tools. Include field names, types, and which fields enable chaining (e.g., marketId, tokenId, walletAddress). Use JSDoc or TypeScript interfaces visible in source.
• Expand tool descriptions to 100-150 chars. Add WHEN to use, prerequisites, and dependencies. E.g., 'Get current trade price for a prediction market token. Use after SEARCH_MARKETS to get tokenId. Returns price in quote token units.'
• Add format constraints to parameter descriptions. E.g., 'walletAddress: BNB Chain wallet address (0x-prefixed hex, 42 chars)' and 'tokenId: Market token identifier (string, typically hex).'
• Document pagination behavior in tool descriptions. E.g., 'Returns up to limit results (default 10, max 20). Use page parameter to fetch additional results. Returns total count for planning.'
• Add error recovery hints to descriptions. E.g., 'If market not found, call SEARCH_MARKETS with keywords. If wallet has no positions, check address is correct and has traded on Opinion.trade.'