Memory-Bank Files Not Automatically Loaded in Context

# Bug Report: Memory-Bank Files Not Automatically Loaded in Context

## Environment
- **Amazon Q Version**: 1.81.0.0
- **IDE**: Visual Studio Code / Visual Studio
- **OS**: Windows

## Issue Description
Memory-bank files created through the Amazon Q chat rules menu are not automatically included in the chat context, despite being stored in the expected `.amazonq\rules\memory-bank\` directory.

## Expected Behavior
Files in `.amazonq\rules\memory-bank\` should be automatically loaded into chat context similar to other rule files in `.amazonq\rules\`.

## Actual Behavior
Memory-bank files are created and stored correctly but are not included in the chat context. The AI assistant cannot access this information without manually reading the files.

## Steps to Reproduce
1. Create memory-bank entries through Amazon Q chat rules menu
2. Verify files are created in `.amazonq\rules\memory-bank\`
3. Start new chat session
4. Ask AI about project-specific information that should be in memory-bank
5. Observe that AI doesn't have access to memory-bank content

## File Structure
```
.amazonq\rules\
├── 00-WORKFLOW.md ✅ Loaded automatically
├── 01-BUILD.md ✅ Loaded automatically
├── 02-FILE-OPS.md ✅ Loaded automatically
├── 03-RESPONSES.md ✅ Loaded automatically
└── memory-bank\
├── guidelines.md ❌ Not loaded automatically
├── product.md ❌ Not loaded automatically
├── structure.md ❌ Not loaded automatically
└── tech.md ❌ Not loaded automatically
```

## Impact
Memory-bank functionality is effectively broken, reducing the value of the rules system for maintaining project context.

## Related Issues
This may be related to the duplicate rule loading bug also present in v1.81.0.0.

Start at the Amazon Q chat rules menu and reproduce the behavior with files in .amazonq\rules\memory-bank\. Compare how the listed root rule files are loaded with how the memory-bank files are handled. Done means memory-bank entries are automatically available in a new chat session without manual file reading.

Written by the indexing model from the issue text.