Server has 11 tools with basic schemas generated via Zod, but significant gaps in descriptions and parameter documentation. Most tools lack detailed descriptions explaining WHEN to use them vs similar tools. Parameter descriptions are minimal or missing entirely. No output schemas documented. Error handling is generic. Tools follow verb_noun naming but descriptions are often under 50 chars, violating the 10-1024 char baseline. Schema generation via zodToJsonSchema is present but underlying Zod definitions not visible in provided code, making full validation impossible.

• Expand all tool descriptions to 100-200 characters. Include WHAT the tool does, WHEN to use it (vs similar tools), and what it returns. E.g., 'Get all accounts following a given account. Use this to discover followers; use get_following() to see who this account follows. Returns account IDs, names, and follow timestamps.'
• Add descriptions to every parameter. For get_account_info, document: 'address: Ethereum address (0x...). identifier: Account username or ID. At least one is required. If both provided, address takes precedence.'
• Document output schemas for each tool. Create a separate section in code comments or a schema file showing expected response structure. E.g., search_atoms returns [{atom_id: string, label: string, type: string}].
• Add error handling with recovery hints. Wrap tool calls in try-catch and return structured errors: {error: 'Account not found', suggestion: 'Try search_account_ids() with a partial name', retryable: false}.
• Consolidate overlapping tools. Merge extract_triples, get_triples_by_ids, and get_triples_with_positions into a single get_triples() tool with optional filters (by_ids, with_positions). Reduce LLM confusion and cognitive load.

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry