← Back to Skills Marketplace
memU: Persistent Memory for 24/7 Agents
by
ProjectSnowWork
· GitHub ↗
· v1.0.0
· MIT-0
502
Downloads
2
Stars
4
Active Installs
1
Versions
Install in OpenClaw
/install memu
Description
Persistent memory infrastructure for 24/7 agents. Replaces flat-file memory with a three-layer architecture (Resource → Memory Item → Memory Category) that r...
Usage Guidance
This Skill is a packaging/documentation wrapper around the open-source memU client (memu-py). Before installing or enabling it: 1) Confirm you are comfortable providing an OpenAI-compatible API key (OPENAI_API_KEY) because the examples and default llm_profiles use it for extraction and retrieval. 2) If you plan to persist memory, secure the PostgreSQL connection string (MEMU_DB_URL) rather than using the example credentials shown. 3) The README mentions an optional memU Cloud (api.memu.so) which would require a separate memU Cloud API key if you choose that mode. 4) Note the project is AGPL-3.0 — if you modify and redistribute components you may have obligations; review the license. 5) As with any third-party Skill, prefer installing memu-py from the official PyPI/repo and verify the upstream memU project (https://github.com/NevaMind-AI/memU) if you need to audit code changes.
Capability Analysis
Type: OpenClaw Skill
Name: memu
Version: 1.0.0
The memu skill bundle is a legitimate integration of the NevaMind AI memU framework, designed to provide persistent, hierarchical memory for agents. The bundle consists of comprehensive documentation (README.md, FAQ.md) and several well-structured Python examples (e.g., example_2_openclaw_integration.py, example_3_production.py) that demonstrate how to store and retrieve structured 'memory items' using either in-memory storage or a PostgreSQL database. The code follows security best practices by utilizing environment variables for sensitive credentials like OPENAI_API_KEY and lacks any signs of data exfiltration, malicious execution, or prompt injection attacks.
Capability Assessment
Purpose & Capability
The name/description (long-term agent memory) aligns with what the skill requires: python3 and an OpenAI-compatible API key for LLM calls. Examples and README all demonstrate memu-py usage and optional PostgreSQL; nothing requested appears unrelated to the stated purpose.
Instruction Scope
SKILL.md and included examples only instruct the agent to call the memu client, configure llm_profiles, run optional Docker for Postgres, and use environment variables for API keys and DB URLs. There are no instructions to read unrelated system files, harvest extra credentials, or exfiltrate data to unknown endpoints. The README mentions an optional hosted memU Cloud API (api.memu.so) — that is optional and would require its own key if used.
Install Mechanism
This is an instruction-only Skill with examples; it recommends pip install memu-py and optionally running a well-known pgvector/postgres Docker image. No arbitrary download URLs, URL shorteners, or extract/install steps are present in the Skill metadata.
Credentials
The only required environment variable declared is OPENAI_API_KEY (also set as primaryEnv), which is proportional for a library that performs LLM calls. Notes: example code shows optional MEMU_DB_URL and example DB connection strings with sample credentials (postgres:postgres) — those are illustrative but remind users to secure real DB credentials. The README also references an optional memU Cloud API key (not declared as required) which is fine as optional, but users should be aware they'd need that key to use the hosted API.
Persistence & Privilege
The skill is not always:true and does not request elevated or persistent platform privileges. It does not modify other skills or system-wide settings. Autonomous invocation is allowed by default (platform standard) but combined with this skill's scoped requests there is no excessive privilege.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install memu - After installation, invoke the skill by name or use
/memu - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release: memU v1.4.0 ClawHub Skill with integration docs, 4 runnable examples, FAQ, and production guide.
Metadata
Frequently Asked Questions
What is memU: Persistent Memory for 24/7 Agents?
Persistent memory infrastructure for 24/7 agents. Replaces flat-file memory with a three-layer architecture (Resource → Memory Item → Memory Category) that r... It is an AI Agent Skill for Claude Code / OpenClaw, with 502 downloads so far.
How do I install memU: Persistent Memory for 24/7 Agents?
Run "/install memu" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is memU: Persistent Memory for 24/7 Agents free?
Yes, memU: Persistent Memory for 24/7 Agents is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does memU: Persistent Memory for 24/7 Agents support?
memU: Persistent Memory for 24/7 Agents is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created memU: Persistent Memory for 24/7 Agents?
It is built and maintained by ProjectSnowWork (@projectsnowwork); the current version is v1.0.0.
More Skills