MWMS AI Skill Installation And Usage Protocol

System: MWMS
Document Type: Protocol
Authority Level: MCR Source Of Truth
Status: Draft For MCR
Primary Location: MCR
Future Operational Destination: HeadOffice Brain, AI Business Systems Brain, AI Manager, AI Employee Router, Brain Room, Course Absorption System, Content Brain, Offer Brain, Sales Brain, Creative Brain, Future AIBS Client Systems
Parent Page: HeadOffice
Owner: Martyn
Developer Boundary: No Development Action Authorized By This Page
Source Of Truth: MCR

Purpose

The purpose of this document is to define the MWMS AI Skill Installation And Usage Protocol.

This protocol establishes how MWMS makes a reusable AI skill available for use, how that skill is triggered, what context it must read, how it should be used safely, and how its output should be reviewed.

MWMS must not treat a skill as active simply because it has been written.

A skill becomes useful only when it is:

defined

stored

assigned

triggerable

context-aware

validated

reviewed

auditable

connected to a handoff path

This protocol exists to prevent AI skills from becoming disconnected documents that sit unused or are used incorrectly.

Without this protocol, MWMS risks:

skills being created but never used

skills being triggered for the wrong task

skills using missing or stale context

skills acting without review

skills duplicating existing procedures

skills ignoring forbidden actions

skills creating outputs with no destination

skills being treated as automation-ready too early

future AIBS client skills being installed without boundaries

The AI Skill Installation And Usage Protocol turns a written skill into an operationally usable MWMS procedure.

Scope

This protocol applies to all MWMS AI skills, reusable AI procedures, AI Employee skill files, skill records, and future AIBS client skill systems.

This includes skills used by:

HeadOffice Brain

AI Business Systems Brain

AI Manager

AI Employee Router

Brain Room

Course Absorption System

Newsletter Intelligence

Opportunity System

Affiliate Brain

Research Brain

Experimentation Brain

Finance Brain

Content Brain

Ads Brain

Creative Brain

Sales Brain

Conversion Brain

Offer Brain

Operations Brain

future AIBS client systems

This protocol applies when MWMS needs to:

install a skill

activate a skill

test a skill

use a skill manually

use a skill in assisted mode

assign a skill to an AI Employee

connect a skill to a context library

define skill trigger phrases

define skill usage boundaries

review skill output

retire or disable a skill

This protocol does not authorize technical development, plugin changes, Supabase changes, WordPress changes, automation wiring, tool access, or M developer action.

Core Definition

AI Skill Installation is the process of making a reusable AI skill available for controlled use inside MWMS.

AI Skill Usage is the process of applying that skill to a real task with the correct input, context, validation, and handoff.

A skill is not installed just because it exists as text.

A skill is installed when MWMS has defined:

where it lives

who owns it

which AI Employee may use it

when it should trigger

what context it needs

what input it accepts

what output it produces

what it must not do

what review is required

where the output goes

what status it holds

Core Principle

The core principle of this protocol is:

A skill must be installed with context, boundaries, and review before it is treated as usable.

MWMS must avoid uncontrolled skill usage.

A skill should make repeated work more consistent, not create new drift.

Installation Readiness Test

Before installing a skill, MWMS must apply the Installation Readiness Test.

Question 1: Is The Skill Clearly Named?

The skill name must describe the work it performs.

Weak name:

Marketing Skill

Strong name:

Context-Grounded Lead Magnet Builder Skill

Question 2: Is The Owning Brain Clear?

Every skill must belong to a Brain.

Question 3: Is The Assigned AI Employee Clear?

The skill should identify which AI Employee uses it.

If no AI Employee owns it, the skill may remain draft.

Question 4: Is The Trigger Clear?

The skill must define when it should be used.

Question 5: Is Required Input Clear?

The skill must define what input it can process.

Question 6: Is Required Context Clear?

The skill must define which context files, standards, or libraries it must read.

Question 7: Is Output Clear?

The skill must define what it produces.

Question 8: Is Validation Clear?

The skill must define how output is checked.

Question 9: Are Forbidden Actions Clear?

The skill must define what it must not do.

Question 10: Is Human Review Clear?

The skill must define whether human review is required.

Possible Outcomes

Install For Manual Use

Install For Assisted Use

Keep As Draft

Needs More Context

Needs More Testing

Merge With Existing Skill

Park For Later

Reject

No installation decision should remain ambiguous.

Skill Installation Statuses

MWMS uses the following installation statuses.

Not Installed

The skill exists as an idea or draft but is not available for use.

Draft Installed

The skill is available for internal testing only.

Manual Use Installed

The skill may be used manually with human review.

Assisted Use Installed

The skill may support repeated AI-assisted workflows with validation.

Controlled Automation Candidate

The skill may be considered for controlled automation later.

Paused

The skill was installed but should not currently be used.

Deprecated

The skill has been replaced.

Retired

The skill is no longer used.

Installation does not mean automation.

Automation requires separate readiness review.

Skill Installation Record

Each installed skill should have an installation record.

Skill Name:

Installation Status:

Owning Brain:

Assigned AI Employee:

Skill Type:

Primary Use Case:

Trigger Phrases:

Required Input:

Required Context:

Related Context Library:

Related Standards:

Required Output:

Validation Requirement:

Human Review Requirement:

Tool Permission Boundary:

Forbidden Actions:

Handoff Destination:

Failure Triggers:

Date Installed:

Installed By:

Last Reviewed:

Next Review:

Notes:

Trigger Phrase Rules

A skill should include clear trigger phrases or trigger conditions.

Trigger phrases help humans and AI Employees know when to use the skill.

Good trigger phrases are specific.

Examples:

Use this when a finished course block needs to be evaluated for MWMS absorption.

Use this when a client lead magnet needs to be built from an approved context library.

Use this when a developer handoff must be prepared for M.

Use this when an existing content folder needs to be scanned for reusable content intelligence.

Weak trigger phrases are vague.

Examples:

Use this for marketing.

Use this for AI.

Use this for content.

Use this when helpful.

Trigger Rule

A skill should trigger because the task requires it, not because the skill exists.

Required Context Rules

A skill must define what context it needs.

Possible context requirements:

current user request

source file

approved Offer Context Library

Client Context Library

Voice Architecture

Objection Library

Proof Library

Retired Language

Compliance Notes

relevant MWMS standards

current save point

developer boundary

human review requirement

Context Rule

A skill without required context is incomplete.

If context is missing, the skill must pause, flag the gap, or operate in draft-only mode.

Skill Usage Modes

MWMS recognizes four skill usage modes.

Mode 1: Manual Use

The AI uses the skill with direct human oversight.

Best for:

new skills

high-risk tasks

MCR page creation

client-facing work

developer handoffs

paid traffic assets

compliance-sensitive work

Mode 2: Assisted Use

The AI uses the skill repeatedly with structured validation.

Best for:

newsletter processing

content briefs

course absorption drafts

offer evaluation drafts

client report drafts

lead magnet drafts

Mode 3: Review Use

The skill is used to check existing output.

Best for:

voice review

retired language review

proof review

context alignment review

MCR structure review

Mode 4: Controlled Automation Candidate

The skill has proven manual or assisted use and may later be considered for automation.

This does not authorize automation.

It only marks future potential.

Skill Usage Workflow

MWMS uses the following workflow when applying a skill.

Step 1: Identify The Task

Confirm what work is being requested.

Step 2: Match The Skill

Determine whether a skill exists for the task.

Step 3: Check Installation Status

Confirm whether the skill is installed and what usage mode is allowed.

Step 4: Confirm Required Input

Check whether the required input is available.

Step 5: Activate Required Context

Load or reference the required context files and standards.

Step 6: Confirm Boundaries

Check forbidden actions, tool permission boundary, human review requirement, and risk level.

Step 7: Run The Skill

Follow the defined procedure.

Step 8: Validate Output

Check output against the skill validation requirement and relevant MWMS standards.

Step 9: Handoff Or Route

Send the output to the correct destination.

Step 10: Capture Learning

If the skill produced a correction, failure, or improvement, record it for audit or refinement.

Skill Usage Rules

Rule 1: Do Not Use Uninstalled Skills Operationally

Draft ideas are not operational skills.

Rule 2: Do Not Use A Skill Without Its Required Context

If context is missing, pause or mark output as draft.

Rule 3: Do Not Let Skills Override Governance

Skills must follow MWMS standards.

Rule 4: Do Not Let Skills Grant Tool Access

Tool permission is separate from skill usage.

Rule 5: Do Not Use Skills For The Wrong Task

A skill should match the task type.

Rule 6: Do Not Treat Assisted Use As Automation

Assisted use still requires validation.

Rule 7: Do Not Skip Human Review On High-Risk Work

Human review gates remain active.

Rule 8: Do Not Leave Outputs Without Destination

Every skill output needs a handoff destination.

Rule 9: Do Not Ignore Repeated Corrections

Repeated correction is a skill audit signal.

Rule 10: Do Not Install Duplicate Skills

Check for existing skills before installing a new one.

Tool Permission Boundary

Skill installation does not create tool permission.

A skill may describe how to perform work.

It may not authorize:

sending email

publishing content

editing WordPress

writing to Supabase

modifying code

accessing client systems

changing dashboards

running automations

deleting files

moving client data

Tool access must be governed by the MWMS AI Tool Permission And Access Framework.

The rule is:

A skill tells the AI how to work. Tool permissions decide what the AI may access.

Human Review Requirements

Human review is required when a skill is used for:

MCR page creation

MCR cleanup

developer handoffs

paid traffic assets

financial analysis

compliance-sensitive content

public-facing sales assets

client-facing reports

future AIBS client systems

tool-enabled workflows

automation candidates

Human review may be lighter for low-risk internal drafts.

Skill Output Requirements

Each skill must produce output in a defined format.

Output may include:

report

page draft

context file

asset draft

skill record

handoff package

validation checklist

audit record

routing decision

parking decision

failure log

Output should not float in chat with no destination.

Skill Failure Triggers

A skill should stop, pause, or escalate when:

required input is missing

required context is missing

source material is weak

output would require invention

duplicate skill risk exists

wrong Brain ownership is unclear

human review is required

tool permission is missing

client context boundary is unclear

compliance risk exists

developer boundary is involved

the output repeatedly fails validation

Failure triggers protect MWMS from pushing weak work forward.

Skill Usage Audit

Installed skills should be reviewed periodically.

Audit questions:

Is this skill being used?

Is it triggering correctly?

Is it producing the expected output?

Is required context clear?

Are forbidden actions clear?

Does output need polishing or rewriting?

Is human review being respected?

Is handoff destination clear?

Does it overlap another skill?

Should it be refreshed, merged, parked, deprecated, or retired?

Client Skill Installation Rules

Future AIBS client skills require stricter control.

Client skill installation must define:

client name

client library used

client approval rules

client data boundary

client tool boundary

client review requirement

client output destination

client privacy level

Client skills must remain isolated.

Client-specific skills must not be installed into general MWMS use unless deliberately generalized and approved.

Common Failure Modes

MWMS must prevent:

skills created but not installed

skills installed without context requirements

skills used for wrong tasks

skills triggering too broadly

skills producing outputs with no destination

skills bypassing human review

skills treated as automation-ready too early

skills granting tool access accidentally

duplicate skills

client skills leaking into general MWMS use

stale skills remaining active

skill outputs becoming source truth without review

Governance Role

HeadOffice owns the MWMS AI Skill Installation And Usage Protocol.

HeadOffice is responsible for:

approving skill installation rules

preventing uncontrolled skill usage

ensuring skills have owning Brains

ensuring skills have trigger conditions

ensuring required context is defined

ensuring human review gates are respected

ensuring tool boundaries are not bypassed

ensuring installed skills are audited

ensuring duplicate skills are avoided

ensuring future client skills remain isolated

Individual Brains may propose and use skills, but HeadOffice governs cross-Brain, high-risk, automation-related, and client-facing skill installation.

Relationship To Other MWMS Standards

This protocol supports and must align with:

MWMS Document Structure Standard

MWMS AI Agent Skill Library Framework

MWMS AI Skill Builder And Audit Protocol

MWMS Source Material To AI Skill Conversion Framework

MWMS AI Agent Memory And Context Framework

MWMS Offer Context Library Standard

MWMS Context Library Governance And Folder Map Standard

MWMS AI Context Activation And Usage Protocol

MWMS AI Brain Audit And Decay Prevention Framework

MWMS AI Brain Readiness Review Checklist

MWMS AI Output Validation Standard

MWMS AI Tool Permission And Access Framework

MWMS AI Employee Role Card Standard

MWMS AI Employee Capability Stack Framework

MWMS AI Employee Handoff Protocol

MWMS Brain Routing Rule

MWMS MCR Promotion To Brain Protocol

MWMS Page Naming Standard

MWMS Architecture Registry

AI Business Systems Brain Canon

This protocol defines how AI skills move from written procedure into controlled use.

Drift Protection

This protocol protects MWMS from:

unused skill documents

uncontrolled skill activation

wrong task triggering

missing context

tool permission overreach

human review bypass

duplicate skills

stale installed skills

client skill leakage

skills becoming automation candidates too early

AI Employees using procedures without validation

Any AI skill used without installation status, trigger, required context, forbidden actions, validation, and handoff destination should be treated as a drift risk.

Architectural Intent

The architectural intent of the MWMS AI Skill Installation And Usage Protocol is to make AI skills operational without making them uncontrolled.

MWMS is building a governed AI workforce.

A workforce needs skills, but those skills need installation rules.

The long-term goal is that every installed skill can answer:

Where does this skill live?

Who owns it?

Who uses it?

When should it trigger?

What input does it require?

What context must it read?

What output does it create?

What must it not do?

What review is required?

Where does output go?

When should it stop?

When should it be audited?

When MWMS can answer these questions consistently, skills become usable procedural intelligence instead of loose prompt documents.

Change Log

v1.0 — Initial Draft

Created the MWMS AI Skill Installation And Usage Protocol as the protocol for installing, activating, triggering, using, validating, reviewing, pausing, and retiring reusable AI skills across MWMS.

This protocol defines installation readiness, installation statuses, installation records, trigger phrase rules, context rules, usage modes, usage workflow, tool boundaries, human review requirements, output requirements, failure triggers, usage audit, client skill installation rules, common failure modes, governance role, drift protection, and architectural intent.

Change Impact Declaration

Pages Created:

MWMS AI Skill Installation And Usage Protocol

Pages Updated:

None

Pages Deprecated:

None

Registries Requiring Update:

MWMS Architecture Registry

HeadOffice Page Registry

AI Business Systems Brain Page Registry

AI Skill Registry

Canon Version Update Required:

No

Change Log Entry Required:

Yes

Employee Impact Check

Employees impacted:

HeadOffice Manager Employee

AI Manager

AI Employee Router

Skill Auditor

Context Library Builder

Course Absorption Agent

Newsletter Signal Extraction Agent

Offer Strategist Employee

Content Planner Employee

Creative Strategist Employee

Sales Strategist Employee

AI Business Systems Architect Employee

Required behaviour updates:

AI Employees must not treat written skill drafts as installed operational skills.

AI Employees must check installation status, trigger conditions, required input, required context, forbidden actions, validation requirements, human review rules, and handoff destination before using a skill.

AI Employees must not allow skills to grant tool permissions or bypass governance.

AI Employees must flag missing context, wrong task fit, duplicate skill risk, or review requirements before producing high-value skill outputs.

AI Employees must keep future client-specific skills isolated unless deliberately generalized and approved.

END MWMS AI SKILL INSTALLATION AND USAGE PROTOCOL v1.0