01PROTOCOL

HORRIS / PROTOCOL

INTENT → SIMULATE →
ASSESS → AUTHORIZE →
EXECUTE.

Horris turns untrusted intent into a bounded execution lifecycle. Each stage narrows authority before capital can move.

LAUNCH TERMINAL →BACK HOME
01INTENT→

Structure the requested action.

Inputs are normalized into a bounded transaction intent before any model output can influence execution.

  • MARKET / ACTION
  • CAPITAL / LIMITS
  • RISK PROFILE
02SIMULATION→

Model the exact state transition.

The route is compiled and simulated against current chain state so Horris can inspect what would actually happen.

  • EXACT CALLDATA
  • LIVE ROUTE STATE
  • NO BROADCAST
03RISK→

Measure adverse outcomes.

Exposure, leverage, loss, slippage, stop distance and policy constraints are evaluated independently of AI confidence.

  • DETERMINISTIC
  • FAIL-CLOSED
  • EXPLAINABLE
04POLICY→

Apply hard authority boundaries.

Allowlists, caps, replay-safe authorization, deadlines and execution constraints decide whether the lifecycle may advance.

  • NO AI AUTHORITY
  • REPLAY SAFE
  • LIMIT ENFORCED
05EXECUTION→

Require explicit wallet authority.

Only approved, simulated and authorized actions become eligible for explicit wallet approval and onchain execution.

  • USER APPROVAL
  • ONCHAIN EVIDENCE
  • PROTECTION AFTER ENTRY
AI AUTHORITYNONE
POLICY AUTHORITYHORRIS
UNVERIFIED ROUTESBLOCKED
EXECUTION MODELFAIL-CLOSED
// HORRIS / SYSTEM PRINCIPLE

Autonomy is permitted only inside explicit boundaries.

Horris is designed so intelligence can propose, but deterministic policy remains the authority that decides whether execution may progress.

[ OPEN CONTROL TERMINAL ]