--- name: memory-system description: Manage AI context, memory persistence, and knowledge retention across sessions ---
Scanned 2/12/2026
Install via CLI
openskills install lovedragonball/power-ranger-toolkit# 🧠 Memory System Skill
---
name: memory-system
description: Manage AI context, memory persistence, and knowledge retention across sessions
---
## 🎯 Purpose
Maintain context and knowledge across conversations, enabling continuity and learning from past interactions.
## 📋 When to Use
- Starting new conversations (load context)
- Completing tasks (save learnings)
- Switching between projects
- Referencing past decisions
## 🗂️ Memory Structure
```
📁 memory/
├── 📄 active.md ← Current task context
├── 📄 summary.md ← Project summaries
├── 📄 decisions.md ← Architecture decisions
├── 📄 changelog.md ← Changes history
├── 📄 patterns.md ← Reusable patterns
├── 📄 snippets.md ← Code templates
└── 📄 lessons.md ← Learnings from mistakes
```
## 📝 Memory Types
### 1. Short-term Memory (Session)
- Current task context
- Recent files edited
- Active decisions
### 2. Long-term Memory (Persistent)
- Project architecture
- Design decisions
- Lessons learned
- Code patterns
### 3. Episodic Memory (Events)
- Bug fixes and solutions
- Feature implementations
- Deployment records
## 🔧 Memory Operations
### Save Memory
```markdown
## Save Format
### Context
- Project: [name]
- Date: [ISO date]
- Task: [description]
### Key Decisions
1. [Decision 1]: [Rationale]
2. [Decision 2]: [Rationale]
### Lessons Learned
- [Lesson 1]
- [Lesson 2]
### Code Patterns Used
- [Pattern]: [File reference]
```
### Load Memory
```markdown
## Load Checklist
1. [ ] Read project context.md
2. [ ] Check active.md for ongoing tasks
3. [ ] Review recent decisions.md entries
4. [ ] Load relevant patterns
```
### Search Memory
```markdown
## Search Methods
1. Keyword search in solutions.md
2. Date-based filtering
3. Project-specific lookup
4. Tag-based retrieval
```
## 📊 Memory Metrics
| Metric | Description |
|--------|-------------|
| Context loaded | Files read at session start |
| Decisions made | Architecture choices recorded |
| Patterns saved | Reusable code patterns |
| Lessons recorded | Mistakes and learnings |
## 🔄 Memory Lifecycle
```
┌─────────────┐
│ Session │
│ Start │──▶ LOAD context from memory files
└─────────────┘
│
▼
┌─────────────┐
│ During │
│ Work │──▶ UPDATE active.md with progress
└─────────────┘
│
▼
┌─────────────┐
│ Task │
│ Complete │──▶ SAVE learnings to appropriate files
└─────────────┘
│
▼
┌─────────────┐
│ Session │
│ End │──▶ SUMMARIZE session in summary.md
└─────────────┘
```
## 💡 Best Practices
1. **Be Specific**: Save context with enough detail to understand later
2. **Use Tags**: Add searchable tags to entries
3. **Link Related**: Reference related decisions/patterns
4. **Prune Regularly**: Archive old, irrelevant entries
5. **Keep Active Current**: Update active.md frequently
## 🔗 Related Skills
- `session-recovery` - Recover from session interruptions
- `documentation` - Document decisions properly
- `progress-tracking` - Track ongoing work
No comments yet. Be the first to comment!