A persistent learning engine that automatically distills conversation insights into long-term memory to prevent repeated mistakes.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install memory-distiller
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install memory-distiller using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is memory-distiller?
memory-distiller acts as the subconscious for your AI agent, bridging the gap between fleeting chat sessions and permanent wisdom. While standard agents often lose context once a session ends, this tool actively monitors interactions to identify valuable corrections, preferences, and insights. By integrating these into your Openclaw Skills workflow, it ensures your agent evolves with every interaction, creating a true learning loop where past lessons inform future actions.
Built on the philosophy that an agent should never be tripped up by the same problem twice, memory-distiller serves as a historian. It reflects on post-conversation data to ensure that when you correct a mistake or state a preference, that information is memorialized. This significantly reduces manual effort in maintaining documentation and improves the long-term utility of your AI workspace.
memory-distiller Use Cases
- Automatically capturing developer preferences for specific coding styles or architectural patterns.
- Recording technical corrections to prevent the agent from repeating command errors or using deprecated syntax.
- Archiving project-specific insights discovered during deep-dive research sessions.
- Building a personalized knowledge base within your Openclaw Skills environment that grows more refined over time.
How memory-distiller Works
- The skill continuously scans every user message for specific trigger signals such as corrections, preferences, or explicit instructions.
- Once a signal is detected, the content is evaluated against a four-stage quality gate: Durability, Generality, Novelty, and Actionability.
- If the insight passes all gates, the skill distills the raw conversation into a structured memory entry.
- The distilled knowledge is automatically written to the appropriate persistent file, such as MEMORY.md or USER.md, without interrupting the user flow.
- In future sessions, the agent references these updated files to provide more accurate, personalized, and context-aware assistance.
memory-distiller Setup
To integrate this autonomous learning capability into your Openclaw Skills setup, run the following command in your terminal:
npx clawhub@latest install memory-distiller
No additional configuration is required. Simply restart your agent session after installation to enable the background distillation process.
memory-distiller Data Schema & Taxonomy
The skill organizes knowledge across three primary Markdown files to ensure high retrieval accuracy and logical organization:
| File | Purpose | Content Type |
|---|---|---|
memory/YYYY-MM-DD.md |
Daily Log | New discoveries, daily lessons, and session-specific insights. |
MEMORY.md |
Long-term Rules | Persistent technical rules, architectural constraints, and general preferences. |
USER.md |
Personal Profile | User-specific personal information and persistent behavioral habits. |
Each auto-learned entry follows a strict schema including a timestamp, the trigger category (Correction, Preference, Insight, or Explicit), and a specific actionable rule that guides future agent behavior.
memory-distiller Advanced Features
- Context-aware multi-file targeting to separate temporary daily logs from global behavioral rules.
- Automated quality filtering system that prevents data bloat by rejecting hypothetical or redundant information.
- Silent background processing for auto-triggered events to maintain a natural conversational experience.
- Privacy-first architecture that automatically excludes sensitive tokens, passwords, and PII from memory files.
- Seamless compatibility with other Openclaw Skills like proactive-agent to create a full real-time and long-term memory stack.
SKILL.md
Loading
Related Openclaw Skills
Prayer Times ID
Automates Indonesian prayer time notifications and Ramadan tracking for the OpenClaw system event ecosystem.
UnifAI Trading Suite
A comprehensive AI toolkit for analyzing prediction markets, social sentiment signals, and economic indicators.
Bio-Reabilita-Z
An AI agent specialized in physical rehabilitation protocols and the management of philanthropic health projects.
Sophie Optimizer
An automated context management utility that monitors token usage, snapshots memory, and resets sessions to maintain peak AI performance.
DOCX Manager
A comprehensive toolkit for programmatically creating, editing, and analyzing Microsoft Word documents with high-fidelity formatting.
HugMe
An automated emotional reset and logic-loop breaking skill that recalibrates AI agents through the HugLLM API when they face frustration or failure.
Featured*
