ArmorIQ SDK

Welcome to the ArmorIQ SDK documentation. This guide will help you integrate and use the ArmorIQ SDK in your applications.

What is ArmorIQ?

ArmorIQ SDK enables you to build intelligent agents that securely execute actions across multiple services (MCPs - Model Context Providers). Think of it as a secure orchestration layer for AI agents.

Key Benefits:

Secure by Design: Cryptographically verified action execution
Intent-Based: Declare what you want to do, not how
Multi-Service: Connect to multiple MCPs with one SDK
Production Ready: Built-in authentication, rate limiting, and monitoring

Core Concepts

Intent-Based Execution

Instead of directly calling services, you declare your intent (what you want to do) upfront. This intent becomes a cryptographically verified contract.

The ArmorIQ Innovation: LLM + Cryptographic Security

ArmorIQ bridges two worlds:

AI Agents that use LLMs to reason and plan dynamically
Zero-Trust Security that cryptographically verifies every action

Traditional Approach:

# Direct calls - no verification
api.call("service1", "action1")
api.call("service2", "action2")
api.call("service3", "action3")  # Could be malicious!

// Direct calls - no verification
api.call('service1', 'action1');
api.call('service2', 'action2');
api.call('service3', 'action3');  // Could be malicious!

ArmorIQ Approach:

# Step 1: Agent captures intent with explicit plan
plan = {
    "goal": "Fetch sales data and analyze Q4 performance",
    "steps": [
        {"action": "fetch_sales", "mcp": "data-mcp", "params": {"quarter": "Q4"}},
        {"action": "analyze", "mcp": "analytics-mcp", "params": {"metrics": ["revenue"]}}
    ]
}

captured_plan = client.capture_plan(
    llm="gpt-4",
    prompt="Fetch sales data and analyze Q4 performance",
    plan=plan  # Explicit plan structure required
)

# Step 2: Get cryptographic proof for the plan
token = client.get_intent_token(captured_plan)

# Step 3: Only declared actions can execute
client.invoke(
    mcp="data-mcp",
    action="fetch_sales",
    intent_token=token,
    params={"quarter": "Q4"}
)   # ✓ Verified (in plan)

client.invoke(
    mcp="analytics-mcp",
    action="analyze",
    intent_token=token,
    params={"metrics": ["revenue"]}
)  # ✓ Verified (in plan)

client.invoke(
    mcp="data-mcp",
    action="delete_all",
    intent_token=token,
    params={}
)    # ✗ Fails - Not in plan!

import { ArmorIQClient } from '@armoriq/sdk';

// Step 1: Agent captures intent with explicit plan
const plan = {
  goal: 'Fetch sales data and analyze Q4 performance',
  steps: [
    { action: 'fetch_sales', mcp: 'data-mcp', params: { quarter: 'Q4' } },
    { action: 'analyze', mcp: 'analytics-mcp', params: { metrics: ['revenue'] } }
  ]
};

const capturedPlan = client.capturePlan(
  'gpt-4',
  'Fetch sales data and analyze Q4 performance',
  plan  // Explicit plan structure required
);

// Step 2: Get cryptographic proof for the plan
const token = await client.getIntentToken(capturedPlan);

// Step 3: Only declared actions can execute
await client.invoke(
  'data-mcp',
  'fetch_sales',
  token,
  { quarter: 'Q4' }
);  // ✓ Verified (in plan)

await client.invoke(
  'analytics-mcp',
  'analyze',
  token,
  { metrics: ['revenue'] }
);  // ✓ Verified (in plan)

await client.invoke(
  'data-mcp',
  'delete_all',
  token,
  {}
);  // ✗ Fails - Not in plan!

Key Insight: You define the exact plan upfront, and every action is cryptographically verified against that plan. This prevents:

Prompt injection attacks (malicious prompts can't execute unplanned actions)
Unauthorized action execution
Plan tampering or modification

The SDK validates your plan structure, then CSRG-IAP creates the cryptographic proof (plan_hash, merkle_root, step_proofs). Each invoke() call is verified against this proof at the Proxy.

Installation - Installation and setup instructions
Client Initialization - How to initialize the ArmorIQ client
Core Methods - Core SDK methods and usage
Data Models - Data structures and models
MCP Directory - MCP directory structure
Error Handling - Error handling patterns
Advanced Usage - Advanced usage patterns
Configuration - Configuration options
Troubleshooting - Common issues and solutions
Best Practices - Best practices and recommendations

ArmorIQ SDK

ArmorIQ SDK

What is ArmorIQ?

Core Concepts

Intent-Based Execution

Getting Started

Installation

Client Initialization

Core Methods

Documentation Sections

On this page