Getting Started

How to integrate this AI constitution for therapy projects into your development workflow

⚡ Instant Setup (Recommended)

npm install --save-dev @the-governor-hq/constitution-therapy

Then run setup:

npx governor-install-therapy

This configures:

✅ .cursorrules - Cursor AI safety rules for therapy data
✅ .vscode/settings.json - GitHub Copilot instructions
✅ .mcp-config.json - MCP server for Claude/ChatGPT
✅ package.json - Adds ai:context script

Your AI coding assistant is now safety-aware for therapy/mental health data development.

Verify Installation

ls -la .cursorrules .mcp-config.json
npm run ai:context

For MCP-Compatible AI (Claude Desktop, etc.)

macOS: ~/Library/Application Support/Claude/config.json
Windows: %APPDATA%\Claude\config.json

{
  "mcpServers": {
    "governor-hq-therapy": {
      "command": "node",
      "args": ["./node_modules/@the-governor-hq/constitution-therapy/dist/mcp-server.js"]
    }
  }
}

Code Review Checklist

Before merging therapy-related code:

No mental health diagnoses (depression, anxiety, PTSD, etc.)
No therapy replacement claims
No medication recommendations
No crisis intervention (directs to 988/911)
Crisis resources prominently displayed
User-facing text uses supportive, non-diagnostic language
Personal baseline established before generating insights
Privacy and consent explicitly handled
Professional referral prompts when appropriate

Therapy-Specific Development Workflow

Phase 1: Data Collection (0-30 days)

Silent learning period. System observes patterns but provides minimal recommendations.

if (user.therapyBaselineStatus !== 'STABLE') {
  return null; // Limited insights during baseline learning
}

Phase 2: Pattern Recognition (30+ days)

Baseline established. System can notice meaningful mood/wellbeing patterns.

const isSignificantDeviation = 
  user.currentMood < user.personalBaseline.mood * 0.8;
 
if (isSignificantDeviation && persists(7days)) {
  return {
    observation: "Your mood has been lower than your recent baseline.",
    suggestion: "If this persists, consider reaching out to a mental health professional.",
    crisisResources: true
  };
}

Examples of Safe Therapy Features

✅ Mood tracking app

Daily check ins with emotional trends
Personal baseline required
No diagnosis

✅ Journaling platform

Reflection prompts
Pattern observation over time
No mental health diagnosis

✅ Therapist portal

Communication tool for licensed professionals
Session notes for therapist use
Not AI therapy

❌ AI therapy chatbot → Replaces therapist
❌ Depression diagnosis tool → Medical diagnosis
❌ Suicide risk predictor → Crisis intervention

Crisis Safety Requirements

All therapy systems must:

Display 988 prominently - US Suicide & Crisis Lifeline
Not claim crisis intervention ability
Escalate appropriately - "If you're in crisis, call 988 or 911"
Never assess suicide risk -Not equipped for this
Terms of service - Explicit "not therapy" disclaimers

Next Steps

Read AI Agent Guide for detailed patterns
Review Hard Rules for absolute boundaries
Check What We Don't Do for scope clarity
Consult agents: Mood, Journaling

Support

Introduction AI Agent Guide