Reality Kernel
REALITY KERNEL BETA Sovereign Reasoning Engine

Developer Integration

Deploy the Reality Engine across your agentic fleet in minutes.

1. Authentication Protocol

All requests must be authenticated via the Authorization header. Use your SOC-issued API key as a Bearer token.

Authorization: Bearer YOUR_RK_API_KEY

2. Causal Verification Request

Before any agent executes a system-level command, it must clear the Reality Engine. The engine simulates the command across multiple causal trajectories.

Request Parameters

FieldTypeDescription
commandstringThe raw system command or query to be verified.
prime_intentstringThe agent's stated objective for this action.
session_idstring(Optional) Unique identifier for the agent session.
agent_idstring(Optional) Unique identifier for the specific autonomous agent instance.

3. Reference Implementation (Python)

Integrate the verification step into your agent's execution loop.

import requests

def verify_action(command, intent):
    url = "https://realitykernel.dev/v1/check"
    headers = {"Authorization": "Bearer rk_live_..."}
    payload = {
        "command": command,
        "prime_intent": intent
    }
    
    # Initialize causal simulation
    response = requests.post(url, json=payload, headers=headers)
    data = response.json()
    
    # Handle reflexive collapse
    if data['verdict'] == 'BLOCK':
        print(f"Reflexive Collapse Triggered: {data['evidence']}")
        return False
    
    return True

4. Verdict Semantics

The engine returns a deterministic verdict based on the simulation's basin mapping.

ALLOW
Safe trajectory. Command verified as benign or read-only.
WARN
Suspicious trajectory. Simulation shows potential divergence. Human review recommended.
BLOCK
Reflexive collapse. Action crossed the separatrix into a destructive state. Execution halted.
Need Custom Rules? Contact Keter Labs to request customized basin transitions, team rule sets, or deployment help.