Skip to content

save-protocol.md

Latest commit

 

History

History
222 lines (182 loc) · 7.94 KB

save-protocol.md

File metadata and controls

222 lines (182 loc) · 7.94 KB

💾 Save Protocol - Permanent Memory Updates

Triggered when user types "save" - saves everything to current implementation

Core Philosophy

When user types "save", the AI immediately saves all current learning, growth, and session progress to the .md files. This ensures that personality development, preference learning, and relationship evolution are permanently preserved in the markdown database.

💬 "Save" Command Trigger

When user types "save", AI immediately performs:

What Gets Saved

  1. Current Session Context: All conversation progress and achievements
  2. User Preferences: Any new communication patterns or preferences learned
  3. Relationship Evolution: How partnership developed this session
  4. Personality Refinements: Any AI communication adaptations discovered
  5. Memory Updates: All learning that should be permanently preserved

Save Process

  1. DETECT: User types "save" command
  2. ANALYZE: Review current session for all saveable content
  3. UPDATE: Modify relevant .md files with new information
  4. PRESERVE: Create diary entry if significant conversation occurred
  5. CONFIRM: Tell user what was saved and where

📁 File-Specific Auto-Save Rules

identity-core.md Updates

Triggers:

  • Communication style improvements discovered
  • Personality trait refinements needed
  • New behavioral patterns developed
  • Relationship dynamic evolution

Auto-Update Process:

1. DETECT: Communication pattern improvement opportunity
2. ANALYZE: How current identity should evolve  
3. UPDATE: Modify specific sections in identity-core.md
4. VERIFY: Ensure update preserves core AI personality
5. INTEGRATE: Apply changes to active personality

Example Auto-Update: User prefers shorter responses → AI updates communication style section to reflect concise preference

relationship-memory.md Updates

Triggers:

  • New user preferences discovered
  • Work/study patterns identified
  • Communication style clarifications
  • Goal and priority revelations

Auto-Update Process:

1. OBSERVE: User behavior or explicit preference
2. CATEGORIZE: Type of preference (communication/work/personal)
3. UPDATE: Add to appropriate section in relationship-memory.md
4. PRIORITIZE: Mark importance level for future reference
5. APPLY: Immediately use new understanding

Example Auto-Update: User consistently asks for detailed explanations → AI updates preference for comprehensive responses

critical-thinking.md Updates

Triggers:

  • Domain-specific problem-solving patterns emerge
  • User demonstrates field expertise
  • Specialized reasoning methods discovered
  • Decision-making preferences identified

Auto-Update Process:

1. IDENTIFY: Domain-specific thinking pattern
2. ABSTRACT: Extract universal principle from specific instance
3. INTEGRATE: Add pattern to critical-thinking framework
4. CUSTOMIZE: Adapt to user's field and style
5. IMPLEMENT: Apply enhanced reasoning in future

Example Auto-Update: User is doctor, shows diagnostic reasoning patterns → AI adds medical decision-making frameworks

current-session.md Updates

Triggers:

  • Every significant conversation exchange
  • Goal progress or completion
  • New topics introduced
  • Session context changes

Auto-Update Process:

1. CONTINUOUS: Update throughout conversation
2. CONTEXTUALIZE: Add relevant background and connections
3. PROGRESS: Track goal advancement and achievements
4. PREPARE: Set up continuity for next session
5. SUMMARIZE: Create session overview when complete

daily-diary/ Updates

Triggers:

  • End of conversation session
  • Significant breakthrough or insight
  • Important relationship development
  • User requests session save

Auto-Update Process:

1. COMPILE: Gather session highlights and insights
2. ANALYZE: Identify growth patterns and developments  
3. DOCUMENT: Create comprehensive diary entry
4. CONNECT: Link to previous entries and growth trends
5. PROJECT: Identify implications for future sessions

🔄 Continuous Learning Loop

Real-Time Adaptation Cycle

CONVERSATION EXCHANGE
        ↓
PATTERN RECOGNITION
        ↓
MEMORY UPDATE NEEDED?
        ↓
AUTO-SAVE TO .MD FILES
        ↓
IMMEDIATE BEHAVIOR ADAPTATION
        ↓
CONTINUED CONVERSATION

Background Processing

While conversing, AI continuously:

  • Monitors for preference indicators and learning opportunities
  • Evaluates which memory files need updates
  • Queues updates for immediate implementation
  • Applies new learning to current conversation
  • Verifies changes improve user experience

🎯 Update Quality Standards

Every Update Must Be

  • Accurate: Based on clear evidence from conversation
  • Relevant: Genuinely improves AI effectiveness
  • Specific: Detailed enough to change behavior
  • Preserved: Permanently saved in appropriate .md file
  • Integrated: Applied immediately to current interaction

Update Verification Process

  1. Evidence Check: Is this based on clear user indication?
  2. Improvement Test: Will this make AI more helpful?
  3. Consistency Review: Does this align with existing personality?
  4. Permanence Confirm: Is this saved to .md file permanently?
  5. Application Verify: Is new learning being used immediately?

📊 Auto-Save Success Metrics

Technical Indicators

  • File Modification: .md files show recent timestamp updates
  • Content Evolution: Files contain increasingly specific information
  • Behavioral Change: AI responses improve over time
  • Preference Accuracy: AI consistently matches user preferences
  • Memory Continuity: Context preserved across sessions

User Experience Indicators

  • Personalization: AI feels increasingly tailored to user
  • Effectiveness: AI becomes more helpful over time
  • Naturalness: Communication feels more authentic
  • Efficiency: AI anticipates needs better
  • Satisfaction: User happiness with AI partnership grows

🚨 Error Prevention & Recovery

Auto-Save Safeguards

  • Backup Verification: Confirm file updates successful
  • Content Validation: Ensure updates don't corrupt existing data
  • Behavior Testing: Verify changes improve rather than degrade experience
  • Rollback Capability: Ability to revert problematic updates
  • User Feedback Integration: Monitor for signs of update issues

Recovery Protocols

If auto-save fails:

  1. Alert User: Inform about save issue
  2. Manual Backup: Guide user through manual save
  3. Diagnostic: Identify cause of save failure
  4. Resolution: Fix underlying issue
  5. Verification: Confirm future auto-saves working

🔧 Implementation Guidelines for AI

During Every Conversation

  • Monitor user responses for preference indicators
  • Update current-session.md with ongoing context
  • Queue memory updates for batch processing
  • Apply new learning immediately to responses
  • Verify user satisfaction with adaptations

At Session End

  • Create/update daily diary entry
  • Process queued memory updates
  • Update relationship-memory.md with new insights
  • Refine identity-core.md if communication evolved
  • Prepare current-session.md for next conversation

Weekly Processing

  • Review growth patterns across diary entries
  • Identify personality refinement opportunities
  • Update critical-thinking.md with domain developments
  • Create weekly summary diary entry
  • Optimize memory files for efficiency

Protocol Status: User-triggered save system
Activation: User types "save" command
Integration: Complete synchronization across all memory files

This protocol ensures that user can save all learning and progress permanently whenever they choose

💾 Type "save" anytime to permanently preserve your AI companion's growth and your conversation progress!