Document Type: Canon
Status: Canon
Authority: MWMS HeadOffice
Applies To: All MWMS brains and AI employees
Version: v1.1
Last Reviewed: 2026-03-15
Parent: MWMS Canon
Purpose
This document defines how Brains communicate, coordinate work, and enforce authority boundaries inside the MWMS ecosystem.
Without a defined interaction protocol, Brains could operate in isolation, duplicate work, or violate governance.
The protocol ensures that all Brains interact through structured, auditable mechanisms rather than informal or uncontrolled exchanges.
This preserves:
• system integrity
• governance enforcement
• operational clarity
• data traceability
Scope
This canon applies to:
• all cross-brain coordination inside MWMS
• all AI employees participating in cross-brain work
• task-based work movement between Brains
• authority-boundary enforcement during collaboration
• logging and audit visibility for cross-brain interaction
This document governs how work moves between Brains and how cross-brain communication must be structured.
It does not govern:
• brain-internal workflow design by itself
• campaign execution by itself
• capital approval by itself
• compliance rulings by themselves
• system architecture wiring by itself
• session-start protocol by itself
Those remain governed by the relevant Brain canons, MWMS System Architecture, MWMS Authority Structure, and How to Start a Session — MWMS Operating Guide.
Definition / Rules
Core Principle
Brains do not communicate through conversation.
Brains communicate through structured tasks and shared system data.
All cross-brain coordination must occur through the MWMS task system.
This ensures every action is:
• recorded
• auditable
• traceable
• reviewable by HeadOffice
MWMS Interaction Model
Brains interact using the following structure:
Requesting Brain
↓
Structured Task Creation
↓
Task Database (Supabase)
↓
Receiving Brain / AI Employee
↓
Execution or Analysis
↓
Result Logged
↓
HeadOffice Visibility
No direct Brain-to-Brain overrides exist.
All coordination flows through the task system.
Task Interaction Types
The MWMS ecosystem recognises several types of cross-brain interaction.
- Intelligence Requests
Used when one Brain requires analysis or research from another Brain.
Example:
Affiliate Brain
→ requests market intelligence
→ Research & Intelligence Brain analyses
→ returns structured intelligence report
Example flow:
Affiliate Brain
↓
Task created: intelligence_request
↓
Research Brain processes task
↓
Results logged to system
- Compliance Verification
Used when operational work must pass integrity or regulatory review.
Example:
Affiliate Brain launches campaign concept
↓
SIT Brain performs compliance validation
↓
Approval or rejection returned
Example flow:
Execution Brain
↓
Task created: compliance_check
↓
SIT Brain reviews
↓
Decision logged
- Financial Oversight
Used when a decision affects capital allocation or risk exposure.
Example:
Media Buying Brain requests campaign budget
Flow:
Media Buying Brain
↓
Task created: budget_request
↓
Finance Brain evaluates capital impact
↓
HeadOffice visibility maintained
- Strategic Escalation
Used when a decision exceeds the authority of a single Brain.
Example:
New product category exploration
Flow:
Originating Brain
↓
Task created: strategic_review
↓
HeadOffice evaluates
↓
Human approval required
- System Integrity Checks
Used to ensure the ecosystem continues operating safely.
Example:
SIT Brain monitors:
• policy violations
• schema inconsistencies
• executor failures
• automation drift
Flow:
SIT Brain
↓
Integrity alert task created
↓
HeadOffice visibility triggered
Authority Enforcement
Brain interaction must always respect the MWMS authority structure.
Authority hierarchy:
Humans (Martyn & M)
↓
HeadOffice
↓
Brains
↓
AI Employees
Rules:
• Brains cannot override HeadOffice
• AI employees cannot override Brains
• Humans retain final authority
Any violation of authority must be escalated.
Brain Autonomy Limits
Brains operate autonomously only within their defined scope.
Autonomy applies to:
• routine operational decisions
• analysis and research tasks
• campaign execution within limits
Autonomy does not apply to:
• capital allocation beyond limits
• new system creation
• governance modification
• structural ecosystem changes
These require escalation.
Task Logging Requirements
Every cross-brain interaction must create a system record containing:
• Task ID
• Originating Brain
• Receiving Brain
• Task Type
• Status
• Timestamp
• Result
This ensures the ecosystem maintains full operational memory.
Failure Handling
If a Brain fails to complete a task:
The task remains visible in the system.
Escalation path:
Receiving Brain
↓
HeadOffice visibility
↓
Human review if necessary
Failure cannot be hidden.
Audit Visibility
HeadOffice maintains visibility over:
• all task creation
• all cross-brain requests
• all execution results
This ensures governance oversight.
Interaction Philosophy
MWMS is not a collection of isolated tools.
It is a coordinated AI organisation.
Brains function like departments in a company.
Departments communicate through structured work requests, not informal conversations.
This ensures:
• accountability
• clarity
• operational discipline
Relationship to Other Canon Documents
This protocol operates alongside:
• MWMS Authority Structure
• MWMS System Architecture
• How to Start a Session — MWMS Operating Guide
Authority defines who controls decisions.
Architecture defines how the system is wired.
Interaction Protocol defines how work moves between Brains.
Mental Model
Authority Structure = Organisational hierarchy
System Architecture = Technical wiring
Interaction Protocol = Department workflow
Together they form the operational framework of MWMS.
Final Principle
A Brain that acts alone becomes chaos.
A Brain that interacts through structure becomes part of a system.
MWMS requires structured interaction.
Final Rule
No cross-brain interaction is valid unless it is task-based, logged, authority-safe, and visible to governance.
Informal coordination is non-compliant.
Drift Protection
The system must prevent:
• informal brain-to-brain requests outside the task system
• cross-brain work occurring without traceable records
• AI employees acting across brain boundaries without authorised routing
• hidden task failure or silent non-completion
• authority breaches being handled casually instead of escalated
• cross-brain coordination drifting into unstructured conversation
Cross-brain interaction must remain structured, logged, and reviewable.
Architectural Intent
MWMS – Brain Interaction Protocol exists to make cross-brain collaboration reliable, auditable, and governance-safe across the ecosystem.
Its role is to ensure that Brains function like coordinated departments inside a structured organisation, with work moving through formal tasks and visible system records rather than informal instruction chains.
Change Log
Version: v1.1
Date: 2026-03-15
Author: MWMS HeadOffice
Change: Rebuilt page to align with the locked MWMS document standard for this cleanup pass. Preserved the original canon intent, task-based interaction model, interaction types, authority hierarchy, autonomy limits, task logging requirements, failure handling, audit visibility, interaction philosophy, related-document logic, mental model, and final principle. Added Document Type, Parent, Scope, Definition / Rules structure, Final Rule, Drift Protection, and Architectural Intent sections.
Version: v1.0
Date: 2026-03-13
Author: MWMS HeadOffice
Change: Initial creation of MWMS – Brain Interaction Protocol defining how Brains communicate, coordinate work, and enforce authority boundaries inside the MWMS ecosystem.
END – MWMS – BRAIN INTERACTION PROTOCOL v1.1