Security Configurator

UAML's Security Configurator gives you centralized control over the security posture of every machine running your AI agents. From firewall rules to disk encryption, manage everything through a unified configuration layer.

Core Components

🧱 Firewall Management

Define and enforce firewall rules across all agent hosts. Whitelist only the ports and services your agents need. Automatic rule generation based on agent configuration — if your agent doesn't use network access, the firewall blocks it by default.

🛡️ Antivirus Integration

Integrates with system antivirus (Windows Defender, ClamAV) to scan memory imports and file attachments before they enter the memory store. Quarantine suspicious content automatically. Real-time protection without manual configuration.

🔐 BitLocker / LUKS Encryption

Enforce full-disk encryption on all agent hosts. UAML verifies encryption status at startup and can block operations if the host disk is unencrypted. Supports BitLocker (Windows) and LUKS (Linux) with status reporting.

Expert Mode

⚡ For Power Users

Expert Mode unlocks advanced security controls: custom firewall rule chains, certificate pinning for agent-to-agent communication, hardware security module (HSM) integration, and fine-grained permission policies. Disabled by default — requires explicit opt-in with acknowledgment.

Security Profiles

Profile	Firewall	AV Scanning	Disk Encryption	Audit Level
Standard	Essential rules	On import	Recommended	Basic
Hardened	Strict whitelist	Real-time	Required	Full
Air-Gapped	All blocked	Real-time	Required	Full + integrity
Expert	Custom rules	Custom	Custom	Custom

Automated Audit Reports

from uaml.security import SecurityConfigurator

config = SecurityConfigurator()

# Check current security posture
status = config.audit()
print(status.score)        # 92/100
print(status.warnings)     # ["SSH key rotation overdue"]

# Generate compliance report
config.generate_report(
    format="pdf",
    standard="iso27001",
    period="2026-Q1",
    output="security-audit-Q1.pdf"
)

# Apply a security profile
config.apply_profile("hardened")

# Verify all hosts comply
compliance = config.verify_fleet()
for host in compliance.non_compliant:
    print(f"⚠️ {host.name}: {host.issues}")
            

Why It Matters

Defense in depth — multiple security layers protect your AI agent's memory
Compliance-ready — automated reports for ISO 27001, SOC 2, and internal audits
Reduced attack surface — principle of least privilege enforced by default
Visibility — know the security status of every agent host at a glance

← Back to UAML

🔧 Security Configurator