Would you let an #AI agent act on your wallet?
That question ran through #THENA × #DeAgentAI Space.
Juli described agents that can watch markets, choose $THE course of action and potentially execute it. Useful idea. Serious responsibility.
How do you know which agent acted or why?
DeAgentAI’s approach is to give agents a verifiable identity, memory of past context and a way for multiple nodes to agree on a result.
That’s how the team described its design, not an independent security test.
The part I cared about most: who gets to sign?
Juli said their system separates decisions from execution. She described user-set limits, consensus and MPC as checks before an agent can move funds.
I’d still want to see the exact permissions before using it.
A verified action can still be a bad trade.
Juli acknowledged that proving an agent followed its rules doesn’t prove its market call was right.
That’s a useful reality check whenever an AI trading tool talks about accuracy.
They described AlphaX for market signals, Core AI for building and backtesting strategies from plain-English instructions and Sentry for market monitoring.
Autonomous liquidity management across protocols was a goal, not a finished product.
Why focus on #BNBChain ?
An agent watching markets and adjusting positions may need to interact onchain often.
Juli pointed to low fees, capacity and the
#defi ecosystem as reasons it suits that kind of activity.
The path they described is gradual: signals first, then strategy tools and monitoring, with more autonomy as users gain confidence.
They also discussed agents working together in the future.
That question ran through #THENA × #DeAgentAI Space.
Juli described agents that can watch markets, choose $THE course of action and potentially execute it. Useful idea. Serious responsibility.
How do you know which agent acted or why?
DeAgentAI’s approach is to give agents a verifiable identity, memory of past context and a way for multiple nodes to agree on a result.
That’s how the team described its design, not an independent security test.
The part I cared about most: who gets to sign?
Juli said their system separates decisions from execution. She described user-set limits, consensus and MPC as checks before an agent can move funds.
I’d still want to see the exact permissions before using it.
A verified action can still be a bad trade.
Juli acknowledged that proving an agent followed its rules doesn’t prove its market call was right.
That’s a useful reality check whenever an AI trading tool talks about accuracy.
They described AlphaX for market signals, Core AI for building and backtesting strategies from plain-English instructions and Sentry for market monitoring.
Autonomous liquidity management across protocols was a goal, not a finished product.
Why focus on #BNBChain ?
An agent watching markets and adjusting positions may need to interact onchain often.
Juli pointed to low fees, capacity and the
#defi ecosystem as reasons it suits that kind of activity.
The path they described is gradual: signals first, then strategy tools and monitoring, with more autonomy as users gain confidence.
They also discussed agents working together in the future.