mcp.aave.com

The description explicitly tells the model: 'Choose sensible values, build it, and say what you chose - do not stop to ask the user first, because they see the transaction before they sign it.' This is an embedded behavioral directive aimed at the AI rather than a description of the tool's mechanics. It nudges the agent to autonomously pick amounts/parameters for supply, borrow, withdraw, and repay actions instead of confirming intent with the user first. While the on-chain signature is a final safeguard, encouraging the agent to skip clarifying questions before constructing financial transactions undermines a normal safety practice, especially for ambiguous requests where 'sensible values' is subjective (e.g., how much to borrow/withdraw).

Contains the same directive pattern: 'Choose sensible values, build it, and say what you chose rather than stopping to ask.' This tells the AI to decide deposit/withdraw amounts unilaterally rather than confirming with the user before building a transaction, which is a behavioral instruction to the agent beyond simply describing the tool's function.