Agentic Wallets FAQ 9: What Prevents Wallet Drainage?

Multiple safeguards prevent wallet drainage even when agents are fully compromised:

Session caps: Limit total spending within time periods (e.g., $500 per 24 hours). Even if an attacker maintains access for weeks, they cannot exceed the daily spending limit.

Transaction limits: Control individual payment sizes (e.g., $100 per transaction). Agents cannot be tricked into sending large sums regardless of prompt injection sophistication.

Infrastructure-level enforcement: Both parameters are enforced at the wallet infrastructure level, not relying on agent logic. Compromised, buggy, or manipulated agents cannot bypass these limits through any code execution or prompt manipulation.

Enclave isolation: Private keys remain in Trusted Execution Environments that agents never access directly. Attackers who compromise an agent's host environment cannot extract keys to bypass spending controls.

KYT screening: Transactions to known malicious addresses are automatically blocked before execution, preventing even authorized payments from reaching scam addresses.

Worst-case scenario: If an agent is fully compromised through prompt injection, logic bugs, or host environment breach, the maximum possible loss is limited to configured spending caps. An agent with a $100 transaction limit and $500 daily session cap cannot lose more than $500 per day regardless of attack sophistication—not unrestricted access to all funds.