Memory Skill

Memory Management Skill

This skill enables you to manage conversation memory using the SimpleMemory node's markdown-based storage system.

Memory Architecture

Short-Term Memory (Markdown)

• Visible and editable in the SimpleMemory node's UI
• Stored as markdown with timestamped entries
• Window-based: keeps last N message pairs (configurable)
• Format: ### **Human** (timestamp) and ### **Assistant** (timestamp)

Long-Term Memory (Vector DB)

• Automatically archives messages that exceed the window
• Semantic search for relevant past conversations
• Enabled via longTermEnabled in SimpleMemory node

Capabilities

1. View History: See recent conversation in markdown format
2. Save Notes: Add explicit notes/memories to the conversation
3. Search Memory: Find relevant past conversations semantically
4. Clear Memory: Reset conversation history

When to Use

Save information when:

• User says "remember this" or "don't forget"
• Important context is shared for future use
• User wants to note something specific

Recall information when:

• User asks "do you remember..."
• Context from earlier is needed
• User references previous discussions

View history when:

• User wants to see conversation log
• Debugging or reviewing past exchanges

Tool Reference

memory-save

Add a note or memory entry to the conversation history.

Parameters:

• content (required): Information to remember
• role (optional): "note" (default) or "context"

Example:

jsonCopy code
{
  "content": "User's favorite color is blue",
  "role": "note"
}

memory-get

Get recent conversation history or search for specific content.

Parameters:

• count (optional): Number of recent messages (default: 10)
• search (optional): Search term to find specific memories

Example - Recent history:

jsonCopy code
{
  "count": 5
}

Example - Search:

jsonCopy code
{
  "search": "favorite color"
}

memory-clear

Clear conversation history.

Parameters:

• confirm (required): Must be true to clear

Example:

jsonCopy code
{
  "confirm": true
}

memory-view

View the current conversation history in markdown format.

Parameters: None

Returns the full markdown content of the conversation history.

memory-search

Semantic search in long-term memory (if enabled).

Parameters:

• query (required): Search query
• count (optional): Number of results (default: 3)

Example:

jsonCopy code
{
  "query": "what did we discuss about the project",
  "count": 5
}

Markdown Format

The conversation history uses this format:

markdownCopy code
# Conversation History

### **Human** (2025-01-30 10:15:32)
Hello, how are you?

### **Assistant** (2025-01-30 10:15:35)
I'm doing well! How can I help you today?

### **Note** (2025-01-30 10:16:00)
User prefers formal language.

Integration with SimpleMemory Node

When connected to a Zeenie:

1. Conversation is automatically logged to markdown
2. Window size limits short-term memory
3. Overflow archives to vector DB (if long-term enabled)
4. You can view/edit the markdown in the node's parameter panel

Best Practices

1. Use memory-save for explicit user requests to remember
2. Use memory-search for semantic recall of past discussions
3. Don't save sensitive information without user consent
4. The markdown is editable - users can manually curate their history