Official skill for Memlink - Universal Memory for AI Agents.
This skill provides comprehensive memory management capabilities for AI agents using the Memlink MCP server. It enables persistent memory across sessions with advanced features like bulk operations, backup/restore, and detailed analytics.
- Memory Management: Create, read, update, and delete memory entries
- Bulk Operations: Delete multiple entries by titles, tags, or patterns
- Search & Filter: Advanced search across all memory entries
- Backup & Restore: Complete backup and restore functionality
- Analytics: Detailed memory statistics and usage analytics
- Safety Features: Dry-run mode for bulk operations
# Via Skills (https://skills.sh)
npx skills rblez/memlink
npx skills https://github.com/rblez/memlink
# Install Memlink globally
npm install -g @rblez/memlink
# Create skill for your agent
memlink agent create <agent-type>Copy the SKILL.md file to your agent's skills directory:
# For Windsurf
cp SKILL.md ~/.codeium/windsurf/skills/memlink/SKILL.md
# For Cursor
cp SKILL.md ~/.cursor/skills/memlink/SKILL.md
# For Claude Desktop
cp SKILL.md ~/.claude/skills/memlink/SKILL.md- memory_read - Read all entries or specific entry by title
- memory_edit - Create or update memory entries
- memory_delete - Delete specific memory entries
- memory_search - Search across all memory entries
- memory_batch - Bulk create/update operations
- bulk_delete - Delete multiple entries by titles, tags, or patterns
{ "method": "tags", "value": "old,deprecated", "dry_run": true }
- backup_create - Create memory backups
- backup_restore - Restore from backup files
- backup_list - List available backups
- backup_delete - Delete specific backups
- backup_cleanup - Clean up old backups
- memory_stats - Get detailed memory statistics
- memory_sync - Validate memory integrity
// Save information to memory
{
"tool": "memory_edit",
"arguments": {
"title": "UserPreferences",
"content": "User prefers concise responses with technical details",
"tags": ["preferences", "style"]
}
}
// Retrieve information
{
"tool": "memory_read",
"arguments": {
"title": "UserPreferences"
}
}// Safe bulk delete with dry run
{
"tool": "bulk_delete",
"arguments": {
"method": "tags",
"value": "test,temporary",
"dry_run": true
}
}// Create backup before major changes
{
"tool": "backup_create",
"arguments": {}
}
// List available backups
{
"tool": "backup_list",
"arguments": {}
}Recommended entry titles:
- UserPreferences - Communication style, language preferences
- ProjectContext - Current project, goals, stack information
- TechStack - Technologies, frameworks, tools being used
- ImportantDecisions - Key architectural or project decisions
- SessionNotes - Temporary session-specific information
- Always backup before bulk operations
- Use dry_run mode for bulk deletions
- Organize entries with consistent tags
- Regular cleanup of outdated entries
- Use search instead of reading all entries for large memories
- ✅ Windsurf (Cascade)
- ✅ Cursor
- ✅ Claude Desktop
- ✅ Devin AI
- ✅ Claude Code
The skill automatically connects to your local Memlink server. Make sure:
- Memlink is installed:
npm install -g @rblez/memlink - Server is running:
memlink serve - MCP is configured for your agent
- Connection refused: Start Memlink server with
memlink serve - Memory not found: Initialize with
memlink init - Permission denied: Check MCP configuration in your agent
# Check Memlink status
memlink status
# List memories
memlink memory list
# Test server
memlink serve --verbose- Fork the repository
- Create feature branch
- Submit pull request
MIT License - see LICENSE file for details.
- Documentation: Memlink Docs
- Issues: GitHub Issues
- Discussions: GitHub Discussions