To get a strategy really running, I care more about whether the boundaries are clear than how clever the model is.
This time, I focused on testing permissions in NeoSoul's NeoTrade:
1. I set position limits, per-trade limits, and a daily order limit for the Agent, and chose Copilot mode, so I have to approve every trade. Skills can adjust entries, exits, and position sizing, but they can't access total capital or the wallet.
2. During operation, several actions were rejected for crossing the boundaries, and the reasons were explained fairly clearly in the logs.
3. This kind of limited autonomy actually makes me more comfortable letting it run. The key to sustained execution is having controllable permissions and a traceable process.
I think it would be even better if boundary warnings were more noticeable.
$DNA
This time, I focused on testing permissions in NeoSoul's NeoTrade:
1. I set position limits, per-trade limits, and a daily order limit for the Agent, and chose Copilot mode, so I have to approve every trade. Skills can adjust entries, exits, and position sizing, but they can't access total capital or the wallet.
2. During operation, several actions were rejected for crossing the boundaries, and the reasons were explained fairly clearly in the logs.
3. This kind of limited autonomy actually makes me more comfortable letting it run. The key to sustained execution is having controllable permissions and a traceable process.
I think it would be even better if boundary warnings were more noticeable.
$DNA