verilink-aiverify-plugin 1.1.0

VeriLinkOS Active Governance Integration for AI Verify

VeriLinkOS AI Verify Plugin

📖 Overview

VeriLinkOS AI Verify Plugin integrates active governance capabilities from VeriLinkOS into the AI Verify testing framework. It provides real-time enforcement, cryptographic evidence, and compliance guardrails for AI testing.

🎯 Why Use This Plugin?

• 🛡️ Active Governance: Prevent unauthorized AI actions before they occur
• 🔐 Cryptographic Receipts: Tamper-evident VAP receipts with Merkle-tree batching
• ✅ AI-BOM Verification: Ensure model integrity and provenance
• 🔑 Trust Tokens: JIT authentication for each test session
• 🧠 Predictive Enforcement: ML-based risk prediction to preemptively block risky tests
• ⚖️ Constitutional AI: Ethical and legal guardrails via MetaRuleSet
• 👥 Multi-Agent Arbitration: Conflict resolution for resource contention

---

🚀 Installation

From PyPI (Recommended)

pip install verilink-aiverify-plugin

Configuration

Set environment variables:

export VERILINK_API_URL=https://api.verilink.os
export VERILINK_API_KEY=your_api_key
export VERILINK_AGENT_ID=your_agent_id
export VERILINK_ORG_ID=your_org_id

Quick Start

import asyncio
from verilink_plugin import VeriLinkGovernancePlugin

async def test_model():
    plugin = VeriLinkGovernancePlugin()
    
    result = await plugin.run_test(
        {"model_id": "model-001", "dataset": "test-data"},
        lambda config: {"passed": True, "artifacts": {}}
    )
    
    print(f"Passed: {result.passed}")
    print(f"Receipt: {result.receipt_id}")
    
    await plugin.cleanup()

asyncio.run(test_model())

License

Apache 2.0