Requirements
- Target platform
- OpenClaw
- Install method
- Manual import
- Extraction
- Extract archive
- Prerequisites
- OpenClaw
- Primary doc
- SKILL.md
Manage agent memory with Membase - a decentralized, encrypted memory backup and restore system. Provides backup, restore, list, diff, status, and cleanup operations for agent memories.
Manage agent memory with Membase - a decentralized, encrypted memory backup and restore system. Provides backup, restore, list, diff, status, and cleanup operations for agent memories.
Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.
I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete.
I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run.
Membase provides secure, decentralized memory storage for AI agents with end-to-end encryption.
Activate this skill when the user: Asks to backup their memories, conversations, or workspace Wants to restore previous memories or conversations Wants to see available backups Asks to compare different backup versions Wants to check backup status Mentions "membase" or "backup memories"
All Membase operations go through a single command: node membase.ts <command> [options] Available commands: backup - Backup memories to Membase restore - Restore memories from a backup list - List all available backups diff - Compare two backups status - Show backup status and statistics cleanup - Clean up old backups
export MEMBASE_ACCOUNT=your-account-address export MEMBASE_SECRET_KEY=your-secret-key export MEMBASE_BACKUP_PASSWORD=your-backup-password export MEMBASE_ENDPOINT=https://testnet.hub.membase.io Check if configured: echo $MEMBASE_ACCOUNT echo $MEMBASE_SECRET_KEY echo $MEMBASE_BACKUP_PASSWORD
Backs up agent memory files (MEMORY.md, memory/**/*.md) to Membase with AES-256-GCM encryption. Usage: node membase.ts backup [options] Options: --password <pwd> or -p <pwd> - Encryption password (required if not in env) --incremental or -i - Only backup changed files since last backup --workspace <path> - Custom workspace directory --no-validate - Skip password strength validation --no-json - Don't output JSON for agent parsing Example conversation: User: "Please backup my memories" You should: Check for MEMBASE_BACKUP_PASSWORD: echo $MEMBASE_BACKUP_PASSWORD If not set, ask: "Please provide a backup password for encryption (at least 12 characters with uppercase, lowercase, and numbers):" Run backup: cd skills/membase node membase.ts backup --password "<password>" Show result to user: [OK] Backup completed Backup ID: backup-2026-02-02T10-30-45-123Z Files: 15 Size: 234 KB [WARNING] Save your backup ID and password securely! Incremental backup (faster): node membase.ts backup --password "<password>" --incremental
Restores memories from a Membase backup. Usage: node membase.ts restore <backup-id> [options] Options: <backup-id> - The backup ID to restore (required) --password <pwd> or -p <pwd> - Decryption password (required if not in env) --no-json - Don't output JSON for agent parsing Example conversation: User: "Restore my memories from backup-2026-02-02T10-30-45-123Z" You should: Check for password: echo $MEMBASE_BACKUP_PASSWORD Run restore: cd skills/membase node membase.ts restore backup-2026-02-02T10-30-45-123Z --password "<password>" Show result: [OK] Restore completed Files restored: 15 Total size: 234 KB Location: ~/.openclaw/workspace/
Lists all available backups for this agent. Usage: node membase.ts list [options] Options: --no-json - Don't output JSON for agent parsing Example conversation: User: "Show me my backups" or "List my backups" You should: cd skills/membase node membase.ts list Output will show: Available backups: ID Timestamp Files Size ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ backup-2026-02-02T10-30-45-123Z 2026-02-02 10:30:45 15 234 KB backup-2026-02-01T15-20-10-456Z 2026-02-01 15:20:10 12 198 KB
Compares two backups to see what changed. Usage: node membase.ts diff <backup-id-1> <backup-id-2> [options] Options: <backup-id-1> - First backup ID (required) <backup-id-2> - Second backup ID (required) --password <pwd> or -p <pwd> - Decryption password (required if not in env) --no-json - Don't output JSON for agent parsing Example conversation: User: "What changed between my last two backups?" You should: Get the two most recent backup IDs: cd skills/membase node membase.ts list Run diff with the two IDs: node membase.ts diff backup-2026-02-02T10-30-45-123Z backup-2026-02-01T15-20-10-456Z --password "<password>" Show result: Added files (2): + memory/conversation-new.md + memory/notes.md Modified files (1): ~ MEMORY.md
Shows backup status and statistics. Usage: node membase.ts status [options] Options: --no-json - Don't output JSON for agent parsing Example conversation: User: "What's my backup status?" or "Check backup status" You should: cd skills/membase node membase.ts status Output shows: [STATS] Backup Status Local: Files: 15 Size: 234 KB Remote: Backups: 10 Configuration: Endpoint: https://testnet.hub.membase.io Agent: my-agent Workspace: ~/.openclaw/workspace
Lists old backups that could be deleted (Membase doesn't support delete API yet). Usage: node membase.ts cleanup [options] Options: --keep-last <n> - Keep last N backups (default: 10) --dry-run - Show what would be deleted without deleting --no-json - Don't output JSON for agent parsing Example conversation: User: "Clean up old backups, keep the last 5" You should: cd skills/membase node membase.ts cleanup --keep-last 5 Note: Will show which backups should be deleted, but user needs to delete manually via Membase Hub UI.
All data is encrypted client-side with AES-256-GCM Password is derived using PBKDF2 with 100,000 iterations Your password never leaves the local machine Membase storage is decentralized and zero-knowledge Only you can decrypt your backups
At least 12 characters Must contain uppercase letters Must contain lowercase letters Must contain numbers Recommended: Use a password manager
If you see "Membase credentials not configured": # User needs to set environment variables: export MEMBASE_ACCOUNT=your-account export MEMBASE_SECRET_KEY=your-key
If you see "Backup password is required": Ask user for password Or suggest setting MEMBASE_BACKUP_PASSWORD env var
If you see "Invalid password" or "Decryption failed": User provided wrong password Ask for correct password
If list shows "No backups found": No backups exist yet Suggest creating first backup
If connection fails: Check internet connection Verify MEMBASE_ENDPOINT is correct Try again later
Always check for password first before asking user Show the backup ID clearly so user can save it Parse JSON output if available (between ---JSON_OUTPUT--- and ---END_JSON---) Be clear about security - emphasize that password is required for restore Suggest incremental backups for speed after first backup Remember backup IDs from list command to help user with restore/diff
# 1. Check status node membase.ts status # 2. First backup (full) node membase.ts backup --password "MySecure123Pass" # 3. Later: incremental backup node membase.ts backup --password "MySecure123Pass" --incremental # 4. List all backups node membase.ts list # 5. Compare recent backups node membase.ts diff backup-id-1 backup-id-2 --password "MySecure123Pass" # 6. Restore if needed node membase.ts restore backup-id-1 --password "MySecure123Pass"
Make sure you're in the skills/membase directory: cd skills/membase pwd # Should show .../skills/membase
The lib folder needs to be linked to compiled source: cd skills/membase ln -sf ../../dist/lib lib
Make membase.ts executable: chmod +x membase.ts
Membase Documentation AgentSkills Specification OpenClaw Skills Guide
Agent frameworks, memory systems, reasoning layers, and model-native orchestration.
Largest current source with strong distribution and engagement signals.