skills.rest
MystenLabs
MystenLabsOFFICIAL
Explore all skills by this author →
Website𝕏@Mysten_LabsGitHub

memwal

Store encrypted AI agent memories on Walrus with on-chain ownership.

83|24|Updated Jul 29, 2025
Search Tags:#delegation#vector-search#walrus#sui#memwal#memory-sdk#encrypted-memory
One-click install
npx skills add https://github.com/MystenLabs/MemWal --skill memwal
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: memwal
Source: https://github.com/MystenLabs/MemWal/tree/main
Command: npx skills add https://github.com/MystenLabs/MemWal --skill memwal

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

MemWal provides a privacy-first memory layer for AI agents by storing memories encrypted on Walrus with on-chain ownership, enabling memory persistence across sessions and devices.

Core Features & Use Cases

  • Persistent, namespace-scoped memory that remains available across restarts and devices
  • End-to-end encryption with SEAL; memory access controlled by owner and delegates
  • Decentralized storage via Walrus with vector search for fast recall
  • Onchain ownership and delegate-key authentication on Sui
  • Use cases include agent memory, cross-app memory sharing, and knowledge retention across sessions

Quick Start

Install the MemWal SDK, initialize MemWal with your delegate key, accountId, serverUrl, and namespace, and call remember to store memories.

Frequently Asked Questions about memwal

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I maintain persistent memory for AI agents across different sessions and devices?▼

Persistent AI memory is maintained by storing encrypted memories on Walrus decentralized storage with on-chain ownership on Sui. This ensures agent knowledge remains available across restarts and devices within a specific owner and namespace boundary.

How does encrypted memory storage and retrieval work for decentralized AI applications?▼

Encrypted memory storage secures data using end-to-end encryption with SEAL and delegate-key authentication on Sui. Retrieval leverages deterministic vector search across the Walrus decentralized storage layer to recall relevant memories quickly.

What do I need to set up persistent AI memory with Walrus and Sui?▼

You need to install the MemWal SDK and operate a running relayer. Configuration requires a delegate key, a MemWalAccount ID on Sui, a serverUrl, and a defined namespace to begin storing and recalling memories.

Can I share AI agent memory across multiple apps while keeping the data secure?▼

Yes, cross-app memory sharing is supported through optional manual or AI middleware integration. Memory access remains strictly controlled by the owner and delegates using on-chain ownership and delegate-key authentication.

How do I recall specific memories from an AI agent's encrypted storage history?▼

You recall specific memories using deterministic vector search. This mechanism queries the encrypted data stored on Walrus to quickly retrieve relevant historical context based on semantic similarity matching.

Related Skills

Explore More Software Engineering Skills→
skalenetworkskalenetwork

ows

Manage secure offline-first multi-chain wallets across EVM, Solana, XRPL, Sui, Bitcoin, Cosmos, Tron, TON, Spark, and Filecoin via CLI or SDK.

Official
Advanced
caobingshengcaobingsheng
1

context-manager

Coordinate dynamic AI context management across memory, tools, and retrieval systems.

Community
Advanced
widnyanawidnyana
4

sui-dev

Route Sui development tasks to Move, TypeScript SDK, frontend, Seal, and Walrus sub-skills.

Community
Intermediate
zphwdzphwd

hummingbot

Develop, configure, and deploy automated crypto trading strategies with Hummingbot.

Community
Advanced