← Back to portfolio

Zero Trust Policy Engine

Status

Prototype

Maturity

Static policy model with decision table

Evidence

Decision logic documented · Interactive simulator planned

Next

Wire up interactive policy toggle UI

Architecture

Actor
Policy Engine
Conditions Check
Allow / Deny Decision
Audit Log

Problem Statement

Traditional perimeter-based access control assumes internal network location implies trust. Modern enterprise environments — remote work, SaaS, multi-cloud, contractor access, service-to-service calls — have no meaningful perimeter. Every access request, regardless of origin, must be evaluated against explicit policy conditions.

Policy Model

Access is granted only when all applicable conditions evaluate to true. Any single condition failure results in an explicit deny with a logged reason. The model covers both human and service account actors, with scope enforcement applying only to non-human identities.

Condition 1

Identity active

Account exists, is not disabled, and is not suspended in the directory

Condition 2

MFA satisfied

A valid second factor was presented and verified within the session

Condition 3

Role valid

The actor holds a role that grants access to the requested resource

Condition 4

Token valid

Presented access token is not expired, not revoked, and signature is valid

Condition 5

Risk acceptable

Risk score for this access context is below the configured threshold

Condition 6

Scope assigned

For service accounts: requested resource is within the assigned OAuth2 scope set

Decision Logic

// Policy evaluation (pseudocode)

ALLOW if:

identity_active == true

AND mfa_satisfied == true (humans only)

AND role_valid == true

AND token_valid == true

AND risk_score < threshold

AND scope_assigned == true (service accounts only)

DENY otherwise — log reason explicitly

Decision Table

ActorRoleResourceMFATokenScopeRiskDecisionReason
AliceEngineerAdmin APINoValidN/ALowDeniedMFA not satisfied
BobOperatorTenant DashboardYesValidN/ALowAllowedAll conditions met
CarolAdminUser DirectoryYesValidN/ALowAllowedFull policy pass
ServiceBotService AccountBilling APIN/AValidMissingLowDeniedRequired scope not assigned
Legacy AppService AccountSCIM APIN/AExpiredAssignedLowDeniedToken expired
ContractorExternalAudit LogsYesValidN/AHighDeniedRisk score too high

Employer Value

Risk reduced

Excessive access, undetected anomalies

Operational value

Consistent, auditable access decisions across all identity types

IAM domain

Zero Trust architecture, policy enforcement, adaptive access

Seniority signal

Thinks in policies, not permissions

Evidence

Policy model documentation

[ Documented above ]

Decision table

[ Documented above ]

Interactive simulator

[ Demo pending ]

GitHub repository

[ Repo pending ]