Design principles
YieldSeeker is built to make autonomous on-chain strategies easier to use without hiding their risks or removing user control.
AI-assisted autonomy
Section titled “AI-assisted autonomy”Agents can evaluate information and carry out a defined strategy without requiring you to approve every routine action. The agent type determines the strategy; the platform does not assume that every agent is optimizing yield.
Simplicity through automation
Section titled “Simplicity through automation”DeFi can require constant attention to markets, protocols, liquidity, incentives, and transactions. YieldSeeker brings those activities into an agent experience while keeping the important decisions and limitations visible.
Constrained execution
Section titled “Constrained execution”An agent should not have unrestricted access to arbitrary blockchain calls. Automated actions pass through the platform’s approved execution paths and the controls described in Security.
Verify, do not trust
Section titled “Verify, do not trust”Where practical, YieldSeeker checks protocol and portfolio information against on-chain state instead of relying only on third-party dashboards or reported metrics. Verification reduces assumptions; it does not remove protocol or market risk.
User sovereignty
Section titled “User sovereignty”Automation should not remove ownership. Users retain the controls provided by their Agent Wallet, including the ability to withdraw or stop participating.
Extensible agent types
Section titled “Extensible agent types”A new agent type should add a new goal and strategy without changing the meaning of the shared wallet and security model. Each type therefore needs its own documentation for supported assets, actions, metrics, and risks.
What these principles do not mean
Section titled “What these principles do not mean”These principles are design goals, not guarantees of profit, safety, liquidity, or uninterrupted service. Read Risks and limitations before using an agent.