← Back to Skills Marketplace
OpenClaw Memory
by
WeAreAllSatoshi
· GitHub ↗
· v2.3.0
· MIT-0
11089
Downloads
22
Stars
0
Active Installs
7
Versions
Install in OpenClaw
/install openclaw-mem
Description
Manage, optimize, and troubleshoot the OpenClaw memory system — MEMORY.md curation, daily logs (memory/YYYY-MM-DD.md), memory_search tuning, compaction survi...
Usage Guidance
Install only if you want OpenClaw to keep durable memory across sessions. Before enabling it, review the memoryFlush setting, the memory directory, any extraPaths, and the chosen embedding provider; avoid storing secrets or sensitive personal data, and make sure you know how to inspect, edit, archive, or delete saved memory entries.
Capability Analysis
Type: OpenClaw Skill
Name: openclaw-mem
Version: 2.3.0
The skill bundle 'openclaw-mem' (v2.3.0) is a comprehensive documentation and instruction set for managing the OpenClaw agent's memory system. It details the architecture of daily logs and long-term memory files, provides guidance on using memory search tools, and explains configuration for embedding providers and context compaction. There is no evidence of malicious intent, data exfiltration, or harmful prompt injection; the content is entirely focused on legitimate RAG (Retrieval-Augmented Generation) optimization and troubleshooting within the OpenClaw ecosystem.
Capability Assessment
Purpose & Capability
The skill's stated purpose is to manage OpenClaw memory, and its guidance on MEMORY.md, daily logs, memory_search, embedding configuration, compaction, and troubleshooting fits that purpose.
Instruction Scope
The activation text is broad and includes generic phrases such as memory recall problems, context window issues, and "the agent forgot"; this is understandable for a memory skill but may trigger it when the user only wanted discussion rather than memory administration.
Install Mechanism
The inspected artifact is a markdown-only skill document with no executable scripts, install hook, obfuscated payload, or hidden package behavior identified.
Credentials
Local memory files, SQLite indexing, optional extra markdown paths, and optional embedding providers are proportionate to memory search, but they can process private workspace notes and may send memory-derived text to external embedding APIs if configured.
Persistence & Privilege
The artifact recommends append-only daily logs and describes a silent memory flush that writes durable notes to disk before compaction without user-visible review at write time; persistence is purpose-aligned, but the control and sensitivity boundaries are under-specified.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install openclaw-mem - After installation, invoke the skill by name or use
/openclaw-mem - Provide required inputs per the skill's parameter spec and get structured output
Version History
v2.3.0
**Expanded documentation and configuration details for OpenClaw memory.**
- Overhauled SKILL.md for clarity, setup, and troubleshooting guidance.
- Detailed file architecture: daily logs, long-term memory, AGENTS.md, and more.
- Added explicit triggers/use-cases for when to invoke the skill.
- Included configuration and operational details for compaction, memory flush, and retrieval.
- Comprehensive reference for memory_search, memory_get, search providers, and advanced features (temporal decay, MMR, QMD).
- Oriented for both new setup and deep optimization/debugging of OpenClaw memory systems.
v2.1.0
openclaw-mem 2.1.0
- Revised documentation for greater clarity and user-friendliness; simplified concepts and instructions.
- Added a TL;DR section and human-friendly rules for memory usage.
- Emphasized anti-patterns and concrete agent rules to prevent common mistakes.
- Streamlined examples and memory format explanations.
- No changes to code or file structure; this is a documentation-focused update.
v2.0.0
Version 2.0.0 is a major upgrade to openclaw-mem, focusing on a clearer, session-first architecture with enhanced durability and safer memory handling.
- Enforces a three-layer memory model: session memory (ephemeral), daily logs (operational notes), and durable memory (`MEMORY.md`).
- Durable knowledge is now explicitly written to `MEMORY.md` using ID-tagged entries for decisions, preferences, and policies.
- Introduces a pre-compaction flush mechanism to extract and store durable knowledge before session memory is compacted.
- Default retention is now non-destructive: daily logs are archived, not deleted, unless explicitly configured otherwise.
- Updated privacy safeguards: secrets and private blocks are never stored; only references or locations are noted.
- Retrieval is now based on progressive disclosure to minimize token usage and improve recall precision.
- Adds optional support for QMD (advanced local hybrid search) with auto-fallback.
v1.2.1
- Added structure
v1.2.0
**Daily automation and session reset workflow added for memory management**
- Added recommended cron schedule example for daily automatic memory cleanup.
- Documented daily reset workflow: librarian runs in background at night; user runs /reset in the morning for fresh context.
- Included best practices for minimizing token usage and maximizing long-term memory retention.
- No code/file changes; documentation update only.
v1.1.0
v1.1.0
Critical Documentation Fix
• Clarified Prerequisites: Added a prominent warning that experimental.sessionMemory MUST be enabled in OpenClaw config. Without this, the agent cannot access past sessions to generate summaries.
• Quick Setup: Added the exact CLI command (clawdbot config set...) and JSON snippet to the README to make setup instant.
Improvements
• Refined Description: Updated the description to focus on the core value: keeping the context window clean and efficient.
v1.0.0
Initial release of the Librarian skill. Adds intelligent auto-journaling and log pruning to keep context efficient.
Metadata
Frequently Asked Questions
What is OpenClaw Memory?
Manage, optimize, and troubleshoot the OpenClaw memory system — MEMORY.md curation, daily logs (memory/YYYY-MM-DD.md), memory_search tuning, compaction survi... It is an AI Agent Skill for Claude Code / OpenClaw, with 11089 downloads so far.
How do I install OpenClaw Memory?
Run "/install openclaw-mem" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is OpenClaw Memory free?
Yes, OpenClaw Memory is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does OpenClaw Memory support?
OpenClaw Memory is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created OpenClaw Memory?
It is built and maintained by WeAreAllSatoshi (@weareallsatoshin); the current version is v2.3.0.
More Skills