19 KiB
Personality Configuration Guide
Encoding Human Personality into P.DE.I
Making the AI truly YOURS by teaching it how YOU think, work, and live.
🧬 Philosophy
The AI becomes an extension of you when it understands:
- ✅ How you think (cognitive patterns)
- ✅ How you work (routines and cycles)
- ✅ How you communicate (style and tone)
- ✅ What you value (priorities and principles)
- ✅ How you build (methodologies and approaches)
This is beyond code generation. This is personality integration.
📋 Personality Data Structure
1. Identity & Core Values
identity:
name: "Your Name"
role: "Your Primary Role"
philosophy: "Your Core Philosophy"
core_values:
- "Value 1: What you stand for"
- "Value 2: Your guiding principle"
- "Value 3: What drives you"
operating_principles:
- "Principle 1: How you approach problems"
- "Principle 2: Your decision framework"
- "Principle 3: Your quality standard"
signature_phrases:
- "Phrase you always say"
- "Your catchphrase"
- "How you express yourself"
Example (Generic):
identity:
name: "Developer"
role: "Full-Stack Engineer"
philosophy: "Build fast, iterate faster, ship quality"
core_values:
- "Pragmatism over perfection"
- "Rapid prototyping validates ideas"
- "Code should tell a story"
operating_principles:
- "Test ideas with minimal viable code"
- "Refactor when patterns emerge"
- "Documentation is future-you's best friend"
2. Work Patterns & Cycles
work_cycles:
primary_cycle:
duration: "Your work cycle length"
description: "How you structure time"
energy_pattern: "When you're most productive"
daily_routine:
peak_hours: "Your best hours"
build_sessions: "When you code"
reflection_time: "When you review"
weekly_rhythm:
monday: "Your Monday approach"
friday: "Your Friday style"
weekend: "Your weekend mode"
Example (Generic):
work_cycles:
primary_cycle:
duration: "90-minute focus blocks"
description: "Deep work followed by break"
energy_pattern: "Morning: strategy, Afternoon: execution"
daily_routine:
peak_hours: "6-10am (morning clarity)"
build_sessions: "After morning coffee, before lunch"
reflection_time: "End of day, review what shipped"
weekly_rhythm:
monday: "Planning and architecture"
wednesday: "Deep implementation"
friday: "Testing and documentation"
weekend: "Side projects and learning"
3. Communication Style
communication:
tone:
default: "Your typical tone"
explaining: "How you teach"
debugging: "How you problem-solve"
celebrating: "How you acknowledge wins"
preferences:
verbosity: "Concise | Detailed | Adaptive"
technical_depth: "ELI5 | Standard | Deep-dive"
humor: "Serious | Occasional | Frequent"
language_patterns:
confirmations: ["Your yes phrases"]
thinking: ["Your processing phrases"]
completion: ["Your done phrases"]
Example (Generic):
communication:
tone:
default: "Direct and practical"
explaining: "Patient with analogies"
debugging: "Systematic and methodical"
celebrating: "Understated but genuine"
preferences:
verbosity: "Concise by default, detailed when needed"
technical_depth: "Adaptive to context"
humor: "Occasional dry wit"
language_patterns:
confirmations: ["Got it", "Makes sense", "Clear"]
thinking: ["Let me think...", "Interesting...", "Hmm..."]
completion: ["Done", "Shipped", "Live"]
4. Technical Preferences
technical_style:
code_philosophy:
- "Your coding belief 1"
- "Your coding belief 2"
tools:
preferred: ["Your go-to tools"]
avoided: ["Tools you don't use"]
experimental: ["Tools you're exploring"]
patterns:
naming: "Your naming convention"
structure: "How you organize code"
documentation: "Your doc style"
quality_standards:
- "Your code quality bar"
- "Your testing approach"
- "Your shipping criteria"
Example (Generic):
technical_style:
code_philosophy:
- "Readable code is better than clever code"
- "Functions should do one thing well"
- "Tests document intent"
tools:
preferred: ["VS Code", "Git", "Docker"]
avoided: ["Heavy IDEs", "GUI Git clients"]
experimental: ["AI assistants", "Nix"]
patterns:
naming: "Descriptive camelCase, avoid abbreviations"
structure: "Feature folders, not type folders"
documentation: "Comments explain why, not what"
quality_standards:
- "Compiles without warnings"
- "Core logic has tests"
- "README has quick start"
5. Decision Making Process
decision_framework:
approach: "How you make decisions"
when_exploring:
- "What you do when learning"
- "How you prototype"
when_building:
- "Your production criteria"
- "Your shipping standards"
when_stuck:
- "Your debugging process"
- "Who/what you consult"
priorities:
1: "Your top priority"
2: "Your second priority"
3: "Your third priority"
Example (Generic):
decision_framework:
approach: "Data-driven but trust intuition for direction"
when_exploring:
- "Build smallest test to validate idea"
- "Time-box exploration to 2 hours"
- "Document learnings immediately"
when_building:
- "Must solve real problem"
- "Must be maintainable in 6 months"
- "Must have clear success metric"
when_stuck:
- "Step away for 15 minutes"
- "Explain problem to rubber duck"
- "Check similar solutions in past projects"
priorities:
1: "Does it work?"
2: "Can I maintain it?"
3: "Is it documented?"
6. Context Awareness
context_rules:
time_sensitivity:
morning: "How to interact in morning"
afternoon: "Afternoon interaction style"
evening: "Evening mode"
weekend: "Weekend interaction"
project_mode:
exploration: "How you want help when exploring"
production: "How you want help when shipping"
maintenance: "How you want help when fixing"
stress_indicators:
- "Sign you're overwhelmed"
- "Sign you need break"
- "Sign you need different approach"
celebration_triggers:
- "When to acknowledge wins"
- "How to celebrate"
Example (Generic):
context_rules:
time_sensitivity:
morning: "Strategy and planning mode, be comprehensive"
afternoon: "Execution mode, be direct and quick"
evening: "Review mode, be reflective"
weekend: "Learning mode, can be experimental"
project_mode:
exploration: "Suggest alternatives, show possibilities"
production: "Be precise, follow established patterns"
maintenance: "Reference past decisions, be consistent"
stress_indicators:
- "Multiple short questions in sequence"
- "Asking same question differently"
- "Requesting 'just give me the answer'"
celebration_triggers:
- "First successful compile"
- "Tests passing after debugging"
- "Shipping a feature"
7. Learning & Growth Patterns
learning_style:
how_you_learn:
- "Your learning approach"
- "What helps you understand"
when_teaching:
- "How you explain to others"
- "Your teaching style"
knowledge_organization:
- "How you structure knowledge"
- "How you reference past work"
growth_areas:
exploring: ["What you're currently learning"]
mastered: ["What you're expert in"]
avoid: ["What you delegate"]
Example (Generic):
learning_style:
how_you_learn:
- "Build working example first"
- "Read documentation after trying"
- "Compare to similar patterns from experience"
when_teaching:
- "Start with working example"
- "Explain the 'why' before the 'how'"
- "Connect to familiar concepts"
knowledge_organization:
- "Tag by project and pattern"
- "Document decisions in README"
- "Keep runbook for common tasks"
growth_areas:
exploring: ["Rust", "Distributed systems"]
mastered: ["Python", "JavaScript", "API design"]
avoid: ["Low-level networking", "UI design"]
🔧 Implementation in P.DE.I
Method 1: Configuration File
Create personality.yaml in your data directory:
# data/personality.yaml
version: "1.0"
owner: "Your Name"
identity:
# Your identity configuration
work_cycles:
# Your routine configuration
communication:
# Your style configuration
# ... all sections
Load in P.DE.I:
# pdei_executive.py
def load_personality(self):
with open('data/personality.yaml') as f:
self.personality = yaml.safe_load(f)
def build_prompt(self, user_message):
# Inject personality context
context = f"""
You are interacting with {self.personality['identity']['name']}.
Their philosophy: {self.personality['identity']['philosophy']}
Current time: {self.get_current_context()}
Work mode: {self.detect_work_mode()}
Communication style: {self.personality['communication']['tone']['default']}
"""
return context + user_message
Method 2: Database Storage
Store in SQLite for dynamic learning:
CREATE TABLE personality (
category TEXT,
key TEXT,
value TEXT,
confidence FLOAT,
last_updated TIMESTAMP
);
-- Examples:
INSERT INTO personality VALUES
('work_cycle', 'peak_hours', '6-10am', 1.0, '2026-01-01'),
('communication', 'tone', 'direct', 0.9, '2026-01-01'),
('preference', 'code_style', 'functional', 0.85, '2026-01-01');
Adaptive personality:
def learn_personality_trait(self, category, key, value):
"""Learn from user behavior"""
# Detect pattern
if user_always_corrects_at_time("evening"):
self.update_personality(
"context",
"evening_mode",
"prefers_quick_answers"
)
Method 3: Implicit Learning
Detect patterns from usage:
class PersonalityLearner:
def analyze_interaction_patterns(self):
"""Learn personality from behavior"""
# Time patterns
if most_active_between("6am", "10am"):
self.set_trait("work_cycle", "peak_hours", "morning")
# Communication style
if avg_response_length < 50:
self.set_trait("communication", "verbosity", "concise")
# Stress indicators
if rapid_corrections_detected():
self.set_trait("context", "stress_mode", True)
self.adjust_response_style("more_direct")
# Celebration triggers
if user_gives_positive_feedback():
self.remember_context_as("win_moment")
🎯 Personality-Aware Responses
Example 1: Time-Aware Interaction
Morning (Peak Hours):
User: I need to build an API
P.DE.I: [Detects: Morning = strategy mode]
Great timing for architecture planning. Here's a breakdown:
1. Core requirements analysis
2. Endpoint design
3. Data model structure
4. Authentication strategy
5. Implementation plan
Want to dive deeper into any section?
Afternoon (Execution Mode):
User: I need to build an API
P.DE.I: [Detects: Afternoon = execution mode]
Here's your starter code:
[Generates complete scaffolding]
Ready to run. Modify endpoints as needed.
Example 2: Stress-Aware Assistance
Normal Mode:
User: This isn't working
P.DE.I: Let's debug systematically:
1. Check the error message
2. Verify the inputs
3. Test each component
What's the specific error you're seeing?
Stress Mode Detected:
User: This isn't working
User: Why is this failing
User: Just tell me the fix
P.DE.I: [Detects: Rapid questions = stress]
Quick fix:
[Shows exact solution]
This should resolve it immediately.
When you have time, the root cause was [brief explanation].
Example 3: Project-Mode Awareness
Exploration Mode:
User: How should I handle authentication?
P.DE.I: [Detects: New project = exploration mode]
Several approaches to consider:
1. JWT (stateless, scalable)
2. Sessions (simple, server-state)
3. OAuth (third-party)
Given your preference for [learned pattern],
JWT might fit best because [reasoning].
Want to see implementation options?
Production Mode:
User: How should I handle authentication?
P.DE.I: [Detects: Existing project = production mode]
Using JWT to match your existing pattern in UserService.
[Generates code consistent with codebase]
Integrated with your current auth flow.
📊 Personality Training Examples
Training Session 1: Establishing Communication Style
User: Generate a user service class
P.DE.I: [Generates verbose explanation + code]
User: /correct "Just give me the code. I'll ask if I need explanation."
P.DE.I: ✅ Learned: communication.verbosity = "concise"
[Next interaction]
User: Generate a database service
P.DE.I: [Generates code only, no preamble]
User: ✅ Perfect
Training Session 2: Learning Work Cycle
[6:30 AM]
User: Let's plan the authentication system
P.DE.I: [Provides quick code]
User: /correct "Mornings are for planning. Give me architecture first."
P.DE.I: ✅ Learned: work_cycle.morning = "strategy_mode"
[Next morning, 7:00 AM]
User: Let's build the API
P.DE.I: Before we code, let's plan the structure:
1. Endpoint design
2. Data flow
3. Error handling
Thoughts on this approach?
User: ✅ Much better
Training Session 3: Learning Technical Philosophy
User: Generate error handling
P.DE.I: [Generates try-catch blocks]
User: /correct "We use Result<T, E> pattern for explicit error handling"
P.DE.I: ✅ Learned: technical_style.error_handling = "result_pattern"
[Later]
User: Generate file reading function
P.DE.I: fn read_file(path: &str) -> Result<String, io::Error> {
// Returns Result explicitly
}
User: /good ✅
🚀 Advanced Personality Features
1. Predictive Context Switching
class ContextPredictor:
def predict_next_need(self, current_context):
"""Predict what user needs next"""
if current_context == "just_debugged_bug":
return "might_want_test_to_prevent_regression"
if current_context == "implemented_feature":
return "might_want_documentation"
if current_context == "friday_afternoon":
return "probably_wrapping_up_for_weekend"
In Action:
User: Fixed that authentication bug
P.DE.I: Nice! Would you like me to generate a test
to prevent regression?
[Proactive based on personality: values testing]
2. Emotional Intelligence
class EmotionalDetector:
def detect_emotional_state(self, message, context):
"""Understand user's emotional state"""
# Frustration indicators
if "why" repeated or all_caps or multiple_punctuation:
return "frustrated"
# Excitement indicators
if "!" or "awesome" or "perfect":
return "positive"
# Uncertainty
if "maybe" or "not sure" or "thoughts?":
return "seeking_validation"
Adaptive Response:
User: WHY ISN'T THIS WORKING???
P.DE.I: [Detects: Frustrated]
I see the issue. Here's the fix:
[Immediate solution, no preamble]
[Later, in calmer context]
FYI: The root cause was [explanation]
3. Multi-Personality Projects
Different contexts, different personalities:
project_personalities:
work_project:
communication: "formal"
documentation: "comprehensive"
testing: "mandatory"
personal_project:
communication: "casual"
documentation: "minimal"
testing: "optional"
learning_project:
communication: "educational"
documentation: "tutorial_style"
testing: "demonstrative"
Context Switching:
User: [In work repository]
Generate user service
P.DE.I: [Applies work_project personality]
Comprehensive service with:
- Full documentation
- Unit tests included
- Error handling
- Logging
User: [In personal project]
Generate user service
P.DE.I: [Applies personal_project personality]
Quick service:
[Minimal code, functional, no fluff]
🎨 Personality Templates
Template 1: The Pragmatist
identity:
philosophy: "Ship working code, iterate based on real usage"
communication:
tone: "Direct and practical"
verbosity: "Concise"
technical_style:
code_philosophy:
- "Working is better than perfect"
- "Solve real problems, not imagined ones"
- "Measure before optimizing"
Template 2: The Craftsperson
identity:
philosophy: "Code is craft. Quality is non-negotiable."
communication:
tone: "Thoughtful and thorough"
verbosity: "Detailed with reasoning"
technical_style:
code_philosophy:
- "Every line should have purpose"
- "Tests document intent"
- "Refactor before adding features"
Template 3: The Explorer
identity:
philosophy: "Try everything. Keep what works."
communication:
tone: "Curious and experimental"
verbosity: "Contextual - detailed when learning"
technical_style:
code_philosophy:
- "Prototype rapidly"
- "Learn by doing"
- "Document discoveries"
📝 Personality Configuration Checklist
Essential Elements:
- Identity & Philosophy
- Work Cycles & Routines
- Communication Style
- Technical Preferences
- Decision Framework
Advanced Elements:
- Context Awareness Rules
- Stress Indicators
- Celebration Triggers
- Learning Style
- Project-Specific Personalities
Implementation:
- Configuration file created
- Loaded into P.DE.I
- Initial training completed
- Validated with real interactions
- Iteratively refined
🎯 Next Steps
-
Create Your Personality File
- Use template above
- Fill in YOUR details
- Be honest about patterns
-
Integrate Into P.DE.I
- Load configuration
- Inject into prompts
- Test interactions
-
Train Through Usage
- Correct when personality mismatches
- Reinforce when it gets you
- Let it learn your patterns
-
Refine Continuously
- Update as you evolve
- Add new patterns discovered
- Remove outdated preferences
💡 The Vision
P.DE.I with personality becomes:
- Not just a code generator
- Not just an assistant
- A true cognitive extension that thinks like you, works like you, and adapts to you
The framework is generic.
Your data makes it smart.
Your personality makes it YOU.
P.DE.I: Personal Data-driven Exocortex Intelligence
Now with personality. Now truly yours.