Policy Service

Plane: Governance Plane

Purpose

Central policy authority. The Policy Service stores, serves, and enforces policies. Every decision-making service in the Intelligence Plane consults the Policy Service before acting. It is the “rules engine” of the platform.

Responsibilities

• Store policy definitions across all three tiers (platform, app, user)

• Serve policy decisions to querying services

• Enforce the policy hierarchy (platform > app > user)

• Scope policies to accounts, organizations, and OUs

• Version and audit policy changes

Policy Tiers

Tier	Author	Override	Examples
Platform	FlexGalaxy.AI	Cannot be overridden	Safety limits, physics constraints
App-defined	Developer	Customizable within bounds	Default failure recovery, scheduling preferences
User-defined	Operator (via AI Agent)	Validated before activation	Custom time-of-day rules, zone restrictions

Key Entities

Entity	Description
Policy	A rule with conditions, actions, and scope
PolicySet	A collection of policies for a domain
Domain Policy Schema	App-registered schema that defines domain vocabulary
Decision	A point-in-time evaluation result

API Surface

Operation	Description
POST /policies	Create a policy
GET /policies?account=&domain=	List policies
POST /evaluate	Evaluate policies for a given context
POST /schemas	Register a domain policy schema
GET /schemas/{domain}	Get schema for a domain

Decision Examples

Querying Service	Question	Policy Answer
Planner	“Which planning strategy for this order type?”	Use priority-first for urgent orders
Scheduler	“Can Robot A run at 2 AM?”	No — time-of-day restriction in effect
Execution Manager	“Contract failed, what now?”	Retry once, then reassign to nearest available
Planner	“Robot battery at 15%, continue?”	No — platform safety limit is 20%

Dependencies

Service	Relationship
DotID	Authentication, account/tenant context
Org Service	Policy scoping to organizations and OUs
Policy Validator	Pre-activation validation

Consumed By

Consumer	Usage
Planner	Planning preferences, constraints
Scheduler	Scheduling priorities, time restrictions
Execution Manager	Failure recovery strategies
AI Policy Agent	Creates policies through conversation
All services	Any service that needs to check “am I allowed to do this?”