Currently, there is a large stock of cryptocurrency bots, but none can prove that they truly follow the rules.

And @MagicNewton $NEWT uses a trusted environment + zk proof to verify automated operations in an encrypted manner, effectively enhancing trust.

For example, stablecoins are the best example: the circulation is $230 billion, but the actual deployment is only 40%, with the remaining portion idle, because many people are reluctant to manually pursue APY returns across protocols.

How to convert intent into automated execution methods without losing trust, instead of relying on bots on Telegram (or X).

Newton is not an application, but an execution layer that provides users with a way to delegate tasks to autonomous agents.

It opens up a series of automations that were previously impossible:

• AI-driven strategies can be adjusted at any time.

• DAO financial bots are responsible for handling expenditures but cannot overspend or deviate from policies.

• Programmable payroll, recurring bills, and even RWA operations are all constrained by zk rules.

• Cross-chain arbitrage executed only when slippage and fee limits are met.

Trust comes from two points:

• TEE - Ensures the code runs as expected.

• ZKP - Ensures it follows your constraints.

Thus, it is no longer "just brutally executing this strategy," but "running this strategy, provided it adheres to the rules."

All conditions must be met, otherwise the system will not settle on-chain.

▫️ Smart accounts + zkPermissions = Programmable delegation.

Users do not just sign transactions; they also define permissions:

What to do, when to do it, under what market conditions, and with what assets.

Then the agent operates within that scope.

For example, the following can be set:

• Only trade when gas is less than 20 gwei.

• End the strategy when my position drops >10%.

• Do not buy or sell a certain token or exceed Y% slippage.

This is a better agent strategy $NEWT.