Critical gaps across all three tools. No input schemas are visible in the source code, tool definitions are inferred from agent.py instructions rather than explicitly registered with JSON Schema. Descriptions exist but lack parameter documentation, format constraints, and output schema specifications. Tools accept no documented parameters, making it impossible for LLMs to understand what inputs they require or how to invoke them. No error handling guidance, no pagination support despite returning 30 days of daily data, and no structured output documentation.

• Create explicit JSON Schema for each tool with input parameters (e.g., get_bitcoin_activity may accept optional date_range, limit, or currency parameters). Register schemas in tools.yaml or via google-adk's tool registration API.
• Add parameter descriptions: document what each input controls, its type, valid range/enum, and format. Example: 'country (enum: USA|China|India|EU|UK), filter macro indicators by region. Defaults to all regions.'
• Document output schemas with typed fields. Example: 'Returns object {date: ISO8601, transaction_count: integer, total_btc_moved: float, avg_fee_satoshis: integer, avg_transaction_btc: float}'.
• Add pagination support: include limit (default 30, max 365) and offset parameters. Return total_days and next_offset in response to enable multi-page queries.
• Expand descriptions to 100 - 200 characters with WHEN to use guidance. Example: 'Call this when the user asks about Bitcoin network activity, whale behavior, or transaction fees. Returns 30 days of daily blockchain metrics.'