agentmesh_drift 3.5.0

Mathematical drift detection library for calculating drift/hallucination scores between outputs

CMVK — Verification Kernel — Public Preview

Part of Agent OS - Kernel-level governance for AI agents

Basic drift detection between model outputs.

CMVK provides lightweight functions for comparing text and embeddings to detect when model outputs diverge. Use it to verify that two models (or two runs of the same model) produce semantically equivalent results.

Installation

pip install agentmesh-drift

Quick Start

from cmvk import verify

score = verify("def add(a, b): return a + b", "def add(x, y): return x + y")
print(f"Drift: {score.drift_score:.3f}")  # 0.0 = identical

Features

• Text comparison with drift scoring (0.0–1.0)
• Embedding and distribution comparison utilities
• Batch verification with aggregation
• Zero external service dependencies (numpy only)

License

MIT License - see LICENSE for details.