Personality Enhancement Configuration

Overview

The ResponsePersonalityEnhancer system allows customization of Piper’s communication style while maintaining accuracy and professionalism. This document covers all configuration options and methods.

Default Configuration

All personality settings have sensible defaults optimized for professional PM work:

Warmth: 0.7 (warm but professional)
Confidence: “contextual” (transparent but not distracting)
Actions: “medium” (helpful without being pushy)
Depth: “balanced” (appropriate for most users)

Configuration Sources (Priority Order)

Database - PersonalityProfile table (persistent, user-specific)
PIPER.user.md - File-based overrides (development/testing)
System Defaults - Fallback values (built-in)

Configuration Structure

Complete Configuration Schema

personality:
  profile:
    warmth_level: 0.7              # 0.0-1.0 (float)
    confidence_style: "contextual"  # enum: numeric|descriptive|contextual|hidden
    action_orientation: "medium"    # enum: high|medium|low
    technical_depth: "balanced"     # enum: detailed|balanced|simplified
  performance:
    max_response_time_ms: 100       # Enhancement timeout (integer)
    cache_ttl_seconds: 300          # Profile cache duration (integer)
    fallback_enabled: true          # Enable graceful degradation (boolean)

Minimal Configuration (PIPER.user.md)

personality:
  profile:
    warmth_level: 0.8
    confidence_style: "hidden"

Configuration Options Details

Warmth Level (0.0 - 1.0)

Controls the enthusiasm and friendliness of responses.

0.0 - 0.2: Very professional, formal, direct
- Example: “Task completed. 3 issues found.”
0.3 - 0.5: Professional with slight warmth
- Example: “Task completed successfully. Found 3 issues for review.”
0.6 - 0.8: Warm and professional (recommended range)
- Example: “Great! Task completed successfully. I found 3 issues that need attention.”
0.9 - 1.0: Very warm and encouraging
- Example: “Excellent work! Task completed perfectly and I found 3 opportunities for improvement!”

Confidence Display Style

Controls how uncertainty and confidence levels are communicated.

“numeric”

Shows specific percentages and numerical confidence indicators.

Example: “Analysis complete (87% confident)”
Example: “Found 5 issues (moderate confidence: 65%)”

“descriptive”

Uses descriptive words for confidence levels.

Example: “Analysis complete (high confidence)”
Example: “Found 5 issues (preliminary analysis)”

“contextual”

Provides context about the confidence assessment.

Example: “Analysis complete (based on recent patterns)”
Example: “Found 5 issues (with current information)”

“hidden”

No confidence indicators shown - clean, streamlined responses.

Example: “Analysis complete”
Example: “Found 5 issues”

Action Orientation

Controls how much actionable guidance is provided.

“high”

Every response includes explicit next steps and recommendations.

Example: “Task completed successfully—ready for the next step! Here’s what I recommend: 1) Review the 3 issues found, 2) Prioritize by impact, 3) Schedule fixes for next sprint.”

“medium”

Actionable guidance when relevant and helpful.

Example: “Task completed successfully. Found 3 issues that need attention—shall I help prioritize them?”

“low”

Minimal action suggestions - focuses on information delivery.

Example: “Task completed successfully. Found 3 issues.”

Technical Depth

Controls the level of technical detail in responses.

“detailed”

Full technical explanations with implementation specifics.

Example: “Database query optimization complete. Reduced query execution time from 847ms to 23ms by adding composite index on (user_id, created_at) and implementing connection pooling with 10-connection limit.”

“balanced”

Right level of detail for most users - technical but accessible.

Example: “Database optimization complete. Query performance improved by 97% through indexing and connection pooling.”

“simplified”

High-level summaries focused on outcomes and business impact.

Example: “Database performance significantly improved. Page load times now under 1 second.”

Environment Variables

Production Configuration

# Personality system settings
PERSONALITY_CACHE_TTL=300                    # Profile cache duration (seconds)
PERSONALITY_MAX_LATENCY_MS=100               # Enhancement timeout (milliseconds)
PERSONALITY_FALLBACK_ENABLED=true           # Enable graceful degradation
PERSONALITY_DEFAULT_WARMTH=0.7               # Default warmth level
PERSONALITY_DEFAULT_CONFIDENCE=contextual    # Default confidence style

# Database settings (if using database profiles)
PERSONALITY_DB_CONNECTION_TIMEOUT=5000       # Database timeout (milliseconds)
PERSONALITY_DB_RETRY_ATTEMPTS=3              # Connection retry attempts

Development/Testing Configuration

# Development overrides
PERSONALITY_CACHE_TTL=10                     # Short cache for testing
PERSONALITY_MAX_LATENCY_MS=200               # Relaxed timeout for debugging
PERSONALITY_FALLBACK_ENABLED=true           # Always enable fallback
PERSONALITY_DEBUG_LOGGING=true               # Enable detailed logging

Configuration Methods

Method 1: Web Interface (Recommended)

URL: http://localhost:8081/personality-preferences

Navigate to personality preferences page
Adjust sliders and dropdowns for desired settings
Click “Save Configuration”
Changes take effect immediately

Advantages:

User-friendly interface
Real-time preview (if available)
Input validation
Persistent storage in database

Method 2: File Configuration (Advanced)

File: ../config/PIPER.user.md

Edit the personality section in PIPER.user.md
Save the file
Restart Piper if needed (depending on configuration)
Changes override database settings

Advantages:

Version controllable
Batch configuration changes
Development/testing friendly
Easy backup and restore

Method 3: API Configuration (Programmatic)

Endpoint: http://localhost:8001/api/personality/profile/default

# Get current configuration
curl -X GET "http://localhost:8001/api/personality/profile/default"

# Update configuration (if endpoint available)
curl -X PUT "http://localhost:8001/api/personality/profile/default" \
  -H "Content-Type: application/json" \
  -d '{
    "warmth_level": 0.8,
    "confidence_style": "contextual",
    "action_orientation": "high",
    "technical_depth": "balanced"
  }'

Configuration Validation

Valid Value Ranges

warmth_level: 0.0 to 1.0 (inclusive)
confidence_style: “numeric”, “descriptive”, “contextual”, “hidden”
action_orientation: “high”, “medium”, “low”
technical_depth: “detailed”, “balanced”, “simplified”

Invalid Configuration Handling

The system handles invalid configurations gracefully:

Out of range values: Clamped to valid range
- warmth_level: 2.0 → 1.0
- warmth_level: -0.5 → 0.0
Invalid enum values: Fall back to defaults
- confidence_style: “invalid” → “contextual”
- action_orientation: “extreme” → “medium”
Missing configuration: Use system defaults
Malformed YAML: Log error and use defaults
Database connection issues: Fall back to file configuration

Performance Considerations

Response Time Impact

Target: <70ms additional latency
Actual: <1ms average enhancement time
Timeout: 100ms maximum (configurable)
Fallback: Original response if timeout exceeded

Caching Strategy

Profile Cache: 300 seconds TTL (configurable)
Enhancement Cache: In-memory for repeated patterns
Database Connections: Pooled and reused
File Watching: Configuration changes detected automatically

Resource Usage

Memory: <10MB additional for personality system
CPU: <5% additional load during enhancement
Network: No additional external calls
Storage: Minimal database storage for profiles

Testing Configuration

Configuration Testing Checklist

Default configuration loads correctly
Web interface saves and loads settings
File configuration overrides work
Invalid values handled gracefully
Performance within acceptable limits
Fallback behavior works when needed

Test Configuration Examples

High Performance Test

personality:
  profile:
    warmth_level: 0.0          # Minimal processing
    confidence_style: "hidden" # No confidence processing
    action_orientation: "low"   # Minimal action processing
  performance:
    max_response_time_ms: 50   # Aggressive timeout
    cache_ttl_seconds: 600     # Long cache

High Engagement Test

personality:
  profile:
    warmth_level: 1.0           # Maximum warmth
    confidence_style: "contextual" # Rich confidence display
    action_orientation: "high"   # Maximum actionability
    technical_depth: "detailed" # Full technical depth

Troubleshooting

Common Configuration Issues

1. Configuration Not Taking Effect

Symptoms: Changes made but responses unchanged Causes:

Cache not cleared
Configuration syntax error
Database connection issues Solutions:
Restart system to clear caches
Validate YAML syntax
Check database connectivity

2. Performance Degradation

Symptoms: Slow response times after enabling personality Causes:

Timeout set too high
Database connection issues
Cache not working Solutions:
Reduce max_response_time_ms
Check database performance
Verify cache TTL settings

3. Inconsistent Personality

Symptoms: Personality varies between interfaces Causes:

Different configuration sources
Cache inconsistency
Interface-specific overrides Solutions:
Check configuration priority order
Clear all caches
Verify consistent configuration across interfaces

Debugging Configuration

Enable Debug Logging

export PERSONALITY_DEBUG_LOGGING=true

Check Configuration Loading

# View current effective configuration
curl http://localhost:8001/api/personality/profile/default | jq '.'

# Check configuration file syntax
python -c "import yaml; print(yaml.safe_load(open('../config/PIPER.user.md')))"

Monitor Performance

# Check response times with personality enabled/disabled
time curl "http://localhost:8001/api/standup?personality=true"
time curl "http://localhost:8001/api/standup?personality=false"

Configuration Guide Version: 1.0 Last Updated: September 11, 2025 Status: Production Ready Next Review: Post-MVP feedback integration