Decide before value moves.

Live reads against the deployed CortexRails Policy on Robinhood Chain testnet: asset state, position, risk, and lifecycle decide whether an action is permitted.

Loading policy console…

Agent-facing intent layer

Agent Intent → Policy → Execution

The agent proposes the action. CortexRails determines whether it's permitted — through the exact same onchain canExecute() read the Policy Console above uses, not a separate simulation.

Offchain agent intentnot yet read from chain

{
  "asset": "TSLA",
  "positionId": "— connect a wallet —",
  "action": "BORROW",
  "amount": "100000",
  "assetOut": "USDG"
}

Amount (USDG)

Connect a wallet to submit an intent.