Single tool 'searchAssets' has a comprehensive schema with 24 parameters, all properly typed and described. Tool name follows verb_noun convention (search + Assets). Description is adequate (127 chars, within 10-1024 range). However, output schema is not documented, the tool returns JSON.stringify() of the Helius API response without specifying what fields the LLM should expect. Error handling returns plain text error messages rather than structured recovery guidance. No tool annotations (readOnlyHint, etc.) despite being a read-only operation. Parameter descriptions are present but some lack actionable constraints (e.g., 'page' has no min/max guidance).

• Document the output schema explicitly. Specify that the response contains 'items' (array of assets), 'total' (count), 'limit', 'page', and per-asset fields like 'id', 'name', 'owner', 'tokenType', 'interface'. Include example structure.
• Add readOnlyHint=true annotation to searchAssets in the tool registration to signal safe retry semantics.
• Enhance error messages with recovery guidance. E.g., 'Invalid ownerAddress format. Expected Solana address (44-44 chars, base58). Got: <value>. Try search_assets with a valid address or call get_user_address first.'
• Add min/max constraints to numeric parameters: page (min=1), limit (min=1, max=100, default=20). Document in parameter descriptions.
• Cap result set at 100 items max and document in tool description: 'Returns up to 100 assets per page. Use page/after parameters to paginate through results.'