LaunchKit · 2026
Back to Skills

This guide walks through setting up the claude-mem plugin on an OpenClaw gateway.

0
0 downloads
by @thedotmack

Setup & Installation

openclaw skills install @thedotmack/claude-mem

Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:

npx clawhub install claude-mem

What This Skill Does

An OpenClaw plugin that gives agents persistent memory across sessions. It records each tool use as a structured observation, writes a MEMORY.md file to the agent's workspace, and optionally streams observations to a messaging channel in real time.

Agents get full session history injected automatically via MEMORY.md without any manual context passing between runs.

When to use it

  • Keeping a coding agent aware of decisions made in past sessions
  • Reviewing what an agent worked on across multiple runs
  • Watching agent tool activity stream to a messaging channel live
  • Scoping memory by project so multiple gateways don't mix observations
  • Auditing agent work history via the worker's built-in viewer UI

Example Workflow

Here's how your AI assistant might use this skill in practice.

INPUT

User asks: 'Continue working on the API client from yesterday'

AGENT
  1. 1Plugin syncs MEMORY.md to agent workspace before session starts
  2. 2Agent reads MEMORY.md and finds prior observations about the API client work
  3. 3Agent continues implementation with full context of previous decisions
  4. 4Each new tool call is recorded as a structured observation by the worker
  5. 5Session ends and worker summarizes progress for the next run
OUTPUT

Agent resumes with accurate prior context; MEMORY.md updated with new observations for future sessions

Requirements

Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.

AI provider API key (e.g., Gemini API key) or Claude Max Plan subscription