uaal-core 0.1.0

Universal AI Authorization Layer (UAAL): policy gating, evidence logging, and Merkle-based verification for autonomous agents

UAAL Core

Universal AI Authorization Layer (UAAL)

UAAL is a lightweight control plane for autonomous AI systems. It enforces policy decisions before actions execute and produces cryptographically verifiable evidence for every decision.

What it does

• Authorizes AI agent actions in real time
• Blocks unsafe or non-compliant actions
• Records immutable evidence per decision
• Generates daily Merkle roots for audit integrity
• Verifies tampering independently

Install

pip install uaal-core

Minimal Usage

from uaal import authorize, emit_evidence

decision = authorize(
    agent="pricing-agent",
    action="update_price",
    payload={"new_price": 120}
)

emit_evidence(decision)

Verify Integrity

from uaal.verify import verify_day
print(verify_day())

Philosophy

UAAL is not an agent framework. It is the seatbelt, black box, and kill switch for AI autonomy.

License

Apache 2.0