setup

A setup guide for connecting Walrus Memory to Claude Code. It uses a local service and a delegate key, which is a separate credential that lets the service act on your behalf.

In plain words
What is it for?
Use it when creating an account, connecting a Sui wallet, logging in, storing credentials locally, or fixing setup problems.
Why use it?
It explains how to install, authenticate, check, and troubleshoot the connection without relying on the browser-based connector flow.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/commandosslabs/walrus-memory-mcp-plugin/setup
Any agent
npx skills add CommandOSSLabs/walrus-memory-mcp-plugin --skill setup
Clone the repo
git clone --depth 1 https://github.com/CommandOSSLabs/walrus-memory-mcp-plugin

Made for: Claude Code, Codex.

Per session 55 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 818 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00055 $0.00818
Opus 5 $0.00028 $0.00409
Sonnet 5 $0.00011 $0.00164
Haiku 4.5 $0.00006 $0.00082

Measured yesterday against content hash c215f157c055, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

setup scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/setup/SKILL.md · 75 lines

How it starts

The opening of the file, as written. The whole thing — 75 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Walrus Memory Setup

Use this skill to help a Claude Code user finish Walrus Memory setup with the published MCP package configured by this plugin.

What This Plugin Uses

This plugin uses the current Walrus Memory MCP package and delegate-key header authentication. It does not depend on the hosted Claude custom-connector OAuth flow.

The plugin starts this MCP server:

{
  "mcpServers": {
    "memwal": {
      "command": "npx",
      "args": ["-y", "@mysten-incubation/memwal-mcp", "--label", "Claude Code Plugin"]
    }
  }
}

Credentials are stored locally at ~/.memwal/credentials.json with file mode 0600.

Setup Flow

  1. Check whether the MCP server is connected.

    • Ask the user to run /mcp if you cannot see MCP tool status.
    • The server name should be memwal.
  2. If tools are available, call memwal_health.

    • If it reports missing credentials, call memwal_login.
    • If memwal_login is not available because the MCP server is not loaded, ask the user to restart Claude Code after installing/enabling the plugin.
  3. During login, the browser opens the Walrus Memory app.

    • The user connects a Sui wallet.
    • If no Walrus Memory account exists, the app walks the user through account creation.
    • The app registers a delegate key labeled Claude Code Plugin.
    • The local MCP package saves delegate-key credentials to ~/.memwal/credentials.json.
  4. After login, verify with memwal_health.

  5. Test end-to-end:

    • Save a harmless setup fact with memwal_remember.
    • Recall it with memwal_recall.

Troubleshooting

  • If the browser flow finishes but Claude Code still says credentials are missing, restart Claude Code, then call memwal_health again. This fallback remains required until the marketplace rollout verifies a published MCP release with live credential reload.
  • If the wallet already has 20 delegate keys, ask the user to open the Walrus Memory dashboard and revoke an unused key.
  • If the user needs a clean login, call memwal_logout, then call memwal_login again.
  • If recall returns nothing for memories that should exist, call memwal_restore with the relevant namespace.
  • If Claude writes MEMORY.md instead of calling memwal_remember, the plugin hooks are not loaded, or Claude's built-in memory is winning. Confirm the plugin is enabled (/plugin) and restart Claude Code. MCP-only is not enough for automatic memory.
  • Show this block, get the user's go-ahead, then merge it into ~/.claude/CLAUDE.md. Create the file if needed. If <!-- memwal:start --> ... <!-- memwal:end --> is already there, replace that span only. Never overwrite the rest of the file.

Read the full file on GitHub · 75 lines

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 75 lines · 55 tokens per session scan A c215f157c055

Subscribe to this mod's changes

setup is a skill published in the GitHub repository CommandOSSLabs/walrus-memory-mcp-plugin (0 stars, last pushed 13d ago), licensed Apache-2.0. It adds 55 tokens to every session and 818 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

sui-frontend

Sui frontend dApp development with @mysten/dapp-kit-react (React) and @mysten/dapp-kit-core (Vue, vanilla JS, other frameworks). Use when building browser apps that connect to Sui wallets, query on-chain data, or execute transactions. Use alongside the sui-ts-sdk skill for PTB construction patterns.

widnyana/eyay-toolkits · 75 tokens

sui-ts-sdk

Sui TypeScript SDK — PTB construction, client setup, transaction execution, and on-chain queries. Use when writing code that interacts with the Sui blockchain via @mysten/sui. These patterns apply in both backend scripts and frontend apps. For frontend-specific setup (dApp Kit, wallet adapters, React hooks), use the…

widnyana/eyay-toolkits · 83 tokens

sui-seal

Seal — decentralized secrets management on Sui. Identity-based encryption with onchain access policies enforced by threshold key servers. Use when building apps that encrypt user data (stored on Walrus, Sui objects, or off-chain) and gate decryption on Move-defined policies. Covers the Move side (sealapprove patterns)…

widnyana/eyay-toolkits · 85 tokens

cover-letter-builder

Build a tailored cover letter from a CV or resume (PDF, plain text, Markdown, LaTeX, or a file path) matched to a job posting. Proactively asks for the user's CV if not provided, then scans the repo as a supplement. Adapts tone, length, and structure to the application channel and company culture. Outputs a polished…

widnyana/eyay-toolkits · 147 tokens

cv-builder

Build or update a Curriculum Vitae (CV) by scanning a repo for the person's background (experience, skills, education, projects — from plain text, Markdown, LaTeX, or any prose file). Outputs a polished, ATS-friendly Markdown CV and optionally a LaTeX or PDF-ready version. Adapts format to target region and remote vs.…

widnyana/eyay-toolkits · 128 tokens

design-method

Design Thinking — the §1–§10 problem-first pipeline: shapes → graph(A) → cardinality → break points (E: ⟳ ↯ ☠) → requirements (R) → boundaries → behavior layers → scope → test layers → verdict. Stack-agnostic. Use when planning, designing, or refactoring any feature or system.

widnyana/eyay-toolkits · 74 tokens