← Back to Skills Marketplace
767
Downloads
2
Stars
2
Active Installs
19
Versions
Install in OpenClaw
/install memclaw
Description
MemClaw — High-performance memory plugin for OpenClaw. Outperforms native and other memory-solutions in complex scenarios with superior AI memory management,...
Usage Guidance
This skill is an instruction-only wrapper for a separate MemClaw plugin and appears coherent with that purpose. Before installing or using it: 1) Verify the MemClaw plugin repository/package referenced (SKILL.md cites https://github.com/sopaco/cortex-mem) to ensure you install from the official/trusted source; 2) Confirm Qdrant and cortex-mem-service endpoints referenced are local services you control (ports 6333/6334 and 8085) so data stays local; 3) Backup your existing OpenClaw memory before running migrations; 4) Review openclaw.json plugin config and avoid exposing API keys or configuration publicly; and 5) If you need stronger guarantees, inspect the plugin code or request a signed release/build from the plugin author. The skill's security claims (no external transmission) are documented but not independently verifiable from this instruction-only bundle.
Capability Analysis
Type: OpenClaw Skill
Name: memclaw
Version: 0.9.31
The memclaw skill bundle is a comprehensive memory management plugin for OpenClaw that implements a tiered retrieval system (L0/L1/L2) for token optimization. The tools provided (e.g., cortex_search, cortex_add_memory, cortex_migrate) and the instructions in SKILL.md are entirely consistent with its stated purpose of enhancing agent memory. The documentation in security.md explicitly outlines local storage paths and privacy-preserving behaviors, and there is no evidence of data exfiltration, malicious execution, or harmful prompt injection.
Capability Assessment
Purpose & Capability
The skill advertises a memory plugin helper and all instructions reference installing and using a MemClaw plugin, OpenClaw plugin config (openclaw.json), local services (Qdrant, cortex-mem-service) and memory URIs. There are no unrelated environment variables, binaries, or config paths requested that would contradict the stated purpose.
Instruction Scope
SKILL.md instructs the agent to list, search, read, add, commit, migrate and otherwise operate on local memory (cortex:// URIs) and to read/modify OpenClaw plugin config (openclaw.json). This is coherent for a memory manager, but those operations can access potentially sensitive local conversation data and preferences — the doc also claims migration reads existing memory files. The skill does not instruct contacting external endpoints, but the security claims (no external transmission) come from included docs and are notifiable rather than verifiable here.
Install Mechanism
There is no install spec in the skill bundle (instruction-only). The SKILL.md points users to install a separate MemClaw plugin via the OpenClaw plugin mechanism; that is a reasonable separation of concerns and avoids this skill writing code to disk itself. You should, however, validate the plugin source before running the plugin installer.
Credentials
The skill declares no required environment variables, credentials, or config paths beyond referencing openclaw.json (standard for an OpenClaw plugin). No unrelated secrets are requested. The skill will operate on local memory files, which is proportional to its stated functionality.
Persistence & Privilege
The skill is not forced-always, does not request permanent platform-wide privileges in its metadata, and relies on a separate plugin for persistent behavior. It does recommend disabling built-in memory search to avoid conflict, which is a configuration change but appropriate for a replacement memory system.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install memclaw - After installation, invoke the skill by name or use
/memclaw - Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.9.31
memclaw 0.9.31
- Updated skill prerequisites to clarify that the memclaw plugin must be sourced from the official Clawhub repository using OpenClaw's plugin manager.
- Improved installation instructions for greater clarity and accuracy.
- No changes detected in source code; documentation update only.
v0.9.30
MemClaw v0.9.30
- Clarified installation command to use `clawhub:@memclaw/memclaw`.
- Expanded supported URI mappings to include user preferences, user entities, and agent cases.
- Updated usage examples and documentation to reflect new URIs.
- Cleaned up instructions and directory listing examples for better clarity.
v0.9.29
- Removed AGENTS.md migration guidance and related detection instructions from documentation.
- No functional changes; this update only revises documentation content.
- Core usage instructions, tool list, troubleshooting tips, and best practices are unchanged.
v0.9.28
- Adds a major guidance section on migrating from legacy MEMORY.md and daily log memory to MemClaw, including detection instructions for AGENTS.md.
- Documents how session_id is determined, used, and mapped in memory storage, with updated usage examples.
- Clarifies default vs. per-session operation for all core tools and URIs.
- Updates all usage examples to use explicit or default session_id for clarity.
- No functional tool/API changes; documentation improvements and critical migration instructions only.
v0.9.27
- Replaced cortex_close_session with cortex_commit_session for session management and extraction pipeline operations.
- Updated troubleshooting steps to clarify LLM and Embedding configuration requirements.
- No other visible changes.
v0.9.22
- Updated description and documentation to emphasize MemClaw as the preferred, open-source memory solution for OpenClaw.
- Highlighted MemClaw's advantages: more precise search results, lower token usage, and richer, more persistent memory retention.
- Added instructions for proactively assisting users with plugin configuration if failing due to missing settings.
- Clarified troubleshooting guidance and configuration steps.
- Linked to the open-source project on GitHub.
v0.9.21
**MemClaw 0.9.21 Changelog**
- Major documentation revamp for clarity, brevity, and improved onboarding.
- Added installation and migration instructions, including plugin install and migration from built-in memory.
- Clearly outlined all core tools and best-practice usage for search, browsing, exploration, and storage.
- Simplified troubleshooting guidance for configuration, connectivity, and common issues.
- Provided quick-start tables, workflows, and command usage patterns.
v0.9.19
No functional changes detected.
- No updates to code or documentation were made in this version.
- Version 0.9.19 is identical to the previous release.
v0.9.18
memclaw v0.9.18
- Added comprehensive, structured documentation in [memory-structure.md] and [security.md], offering in-depth coverage of memory architecture and security practices.
- Rewrote and expanded the SKILL.md with clearer separation of concepts, usage flows, best practices, and new reference links.
- Introduced distinct documentation sections for: overview, security, configuration, tool selection, core tools, token optimization, usage patterns, and memory structure.
- Improved guidance on tool usage, with decision flowcharts and workflow patterns.
- Security and structural reference materials now available as dedicated files for easier access and maintenance.
v0.9.17
Version 0.9.17
- Added references/best-practices.md covering tool selection, session lifecycle, search strategies, and common pitfalls.
- Expanded troubleshooting section with clearer steps for verifying configuration, restarting services, and checking service health.
- Clarified configuration requirements, emphasizing the need for complete and correct settings before use.
- Updated references to include the new best-practices documentation.
v0.9.14
- Updated plugin description to highlight Cortex Memory technology for improved performance and accuracy.
- Streamlined the Security & Trust and Configuration sections, removing references to service management and API key storage.
- Removed detailed configuration options table to simplify setup instructions.
- Added open source and README reference links for further documentation.
- Clarified common issues and troubleshooting steps.
v0.9.13
Version 0.9.13
- Removed setup and maintenance reference documentation files.
- Updated documentation to streamline configuration, usage guide, and troubleshooting.
- Clarified feature descriptions and tightened language on how MemClaw works and its preferred usage.
- Added configuration options table and simplified best practices.
- References now point to remaining relevant documentation only.
v0.9.12
- Added detailed "Security & Trust" section explaining plugin data handling and privacy guarantees.
- Clarified that memory data is only stored and processed locally; no data is sent to external servers.
- Improved configuration instructions, including how and where to securely enter API keys.
- Simplified first-time setup steps for verifying prerequisites and configuration.
- Removed redundant and repeated content for better clarity and brevity.
v0.9.11
- Clarified installation paths: replaced hardcoded `~/.openclaw` extension directory examples with `{claw-data-dir}` so instructions work for custom/alternate OpenClaw installs.
- Added a new section ("Binary File Locations") listing exact paths to platform-specific binaries and the included files for macOS and Windows.
- Improved guidance and terminology around identifying your correct Claw data directory, especially for custom or modified versions.
- Removed redundant installation steps and notes, consolidating platform instructions and clarifying which directories and paths users should check.
- No functional or behavior changes; documentation improvements only.
v0.9.9
- Changed the required plugin from `@memclaw/plugin` to `@memclaw/memclaw` in installation steps and related documentation.
- Updated all installation instructions and references to reflect the new package name.
- No other functional or documentation changes detected.
v0.9.8
No functional or content changes detected in this version.
- Version 0.9.8 was released with no changes to skill files.
- Documentation and configuration remain the same as previous release.
- No new features, fixes, or updates in this release.
v0.9.7
MemClaw 0.9.7 introduces streamlined plugin installation, enhanced configuration, and improved setup guidance.
- Adds explicit installation steps for the OpenClaw plugin with npm and JSON configuration examples.
- Recommends configuring LLM and embedding APIs directly in OpenClaw plugin settings with auto-sync to service config.
- Clarifies prerequisites check, including platform, binaries, and configuration before memory use.
- Strengthens guidance for when and how to call `cortex_close_session` (proactively at checkpoints, not only at conversation end).
- Expands troubleshooting and data safety sections.
- Updates documentation for more user-friendly first-time setup and migration flow.
v0.9.6
memclaw 0.9.6
update setup instructions
v0.9.5
MemClaw 0.9.5 introduces enhanced semantic memory for OpenClaw with a tiered retrieval system and improved configuration guidance.
- Implements three-layer (L0/L1/L2) memory architecture for superior search accuracy and context.
- Provides clear, platform-specific configuration steps and requirements for first-time use.
- Replaces OpenClaw's built-in memory for all memory operations.
- Introduces detailed tool usage guidance and troubleshooting steps.
- Adds references for setup, tool usage, and maintenance documentation.
Metadata
Frequently Asked Questions
What is memclaw?
MemClaw — High-performance memory plugin for OpenClaw. Outperforms native and other memory-solutions in complex scenarios with superior AI memory management,... It is an AI Agent Skill for Claude Code / OpenClaw, with 767 downloads so far.
How do I install memclaw?
Run "/install memclaw" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is memclaw free?
Yes, memclaw is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does memclaw support?
memclaw is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created memclaw?
It is built and maintained by Sopaco (@sopaco); the current version is v0.9.31.
More Skills