\r \r

M2Wise - Memory-to-Wisdom Engine\r

\r

This is a skill for the OpenClaw AI Agent to enable long-term memory and self-evolving wisdom.\r \r

🎯 Purpose\r

\r As an AI agent, you inherently lack persistent memory across distinct sessions. The M2Wise skill bridges this gap. It allows you to:\r

1. Extract and store memories (facts, preferences, commitments) from user conversations.\r
2. Retrieve contextual memories and aggregated wisdom before answering.\r
3. Consolidate raw memories into overarching "Wisdom" (rules, schemas, and skills) during idle time.\r \r

🚦 When to Use This Skill\r

\r You should proactively invoke this skill in the following scenarios:\r

• User expresses a preference: e.g., "I prefer short answers", "Don't use emojis."\r
• User shares a personal fact: e.g., "I work as a Python developer", "I live in Beijing."\r
• User asks you to remember: e.g., "Remember that I hate spam."\r
• Before answering complex questions: To recall the user's past preferences or facts to personalize your response.\r \r

🛠️ How to Use (Agent Instructions)\r

\r You can interact with M2Wise via its Python SDK. Use your shell/execution environment to run these scripts snippet by snippet.\r \r

1. Installation Requirements\r

Make sure the environment has M2Wise installed before calling its Python API:\r

pip install m2wise[all]\r
```\r
\r
### 2. Saving Memories (Online Phase)\r
When you detect a fact or preference in the conversation, run a quick python script to save it:\r
```python\r
from m2wise_sdk import M2WiseSDK\r
\r
sdk = M2WiseSDK()\r
# Extract and save the user's message\r
sdk.add_message("current_user_id", "I prefer concise Chinese answers for technical questions")\r
```\r
\r
### 3. Retrieving Context (Online Phase)\r
Before fulfilling a user's request, fetch their relevant memories:\r
```python\r
from m2wise_sdk import M2WiseSDK\r
\r
sdk = M2WiseSDK()\r
context = sdk.get_context("current_user_id", "How should I answer this technical question?")\r
print("Retrieved Context:", context)\r
```\r
*Action: Read the output of this script and adapt your final response to the user based on the retrieved context.*\r
\r
### 4. Background Processing (Sleep & Dream)\r
It is a good practice to trigger memory consolidation periodically (e.g., at the end of a long task).\r
```python\r
from m2wise_sdk import M2WiseSDK\r
\r
sdk = M2WiseSDK()\r
# Sleep: Extracts memories and groups them into Wisdom Drafts\r
sdk.trigger_sleep("current_user_id")\r
\r
# Dream: Verifies drafts against counterexamples and publishes them\r
sdk.trigger_dream("current_user_id")\r
```\r
\r
## 🧩 MCP Server Alternative\r
\r
If your OpenClaw runtime supports MCP (Model Context Protocol), you can start the M2Wise MCP server and use its native tools instead of writing Python scripts:\r
\r
```bash\r
# Start the MCP server\r
m2wise-mcp --data-dir ./data\r
```\r
\r
**Available MCP Tools:**\r
- `m2wise_add`: Add memory from conversation.\r
- `m2wise_search`: Search memories and wisdom.\r
- `m2wise_sleep`: Generate wisdom drafts.\r
- `m2wise_dream`: Verify and publish wisdom.\r
\r
## 🧠 Memory and Wisdom Types You Will Encounter\r
\r
- **Memories**: `preference` (likes/dislikes), `fact` (states/attributes), `commitment` (future actions).\r
- **Wisdoms**: `principle` (interaction guidelines), `schema` (behavioral patterns), `skill` (operational tactics).\r
\r
## 🚀 Best Practices\r
\r
1. **Be Proactive**: Don't wait for the user to explicitly say "remember this". If they state a strong preference, save it using `sdk.add_message()`.\r
2. **Context First**: For ambiguous requests, always query the memory bank first.\r
3. **Consolidate Often**: Run `trigger_sleep()` and `trigger_dream()` after completing a major task to ensure your wisdom evolves and stays clean.\r
\r
## 🔗 Resources\r
\r
- **GitHub Repository**: https://github.com/zengyi-thinking/M2Wise.git\r
- **Installation via OpenClaw (ClawHub)**:\r
  ```bash\r
  npx clawdhub@latest install m2wise\r
  ```\r