memory-system
自研三层记忆恢复系统。解决session重启后"忘记一切"的问题，提供永久记忆、今日记忆、临时记忆的完整架构。
Install via CLI (Recommended)
clawhub install openclaw/skills/skills/daoistbro/memory-systemWhat This Skill Does
The memory-system is a robust, three-tiered hierarchical memory management architecture designed to eliminate the 'amnesia' issue often experienced by AI agents when sessions reset. By segmenting data into Permanent (identity, stacks), Daily (logs, session history), and Temporary (live context) storage, the skill provides a consistent and persistent personality and context-retention framework. It functions as an automated journaling system that saves key decisions, technical habits, and evolving preferences to local markdown files, ensuring that the agent remains 'woke' even after complete session reboots.
Installation
To install this skill, run the following command in your terminal: clawhub install openclaw/skills/skills/daoistbro/memory-system
Use Cases
- Project Continuity: Retain architectural decisions and technical constraints across multi-day coding tasks.
- Agent Personalization: Maintain user preferences and work-style rapport over weeks or months.
- Auditability: Review daily logs to understand how and why specific technical decisions were made.
- Disaster Recovery: Quickly reconstruct the state of a complex project after an unexpected environment wipe.
Example Prompts
- 'Memory system, please summarize what we decided regarding the database schema in yesterday's session.'
- 'Update my permanent memory: I now prefer to use Tailwind CSS for all frontend components in this project.'
- 'I'm starting a new session; pull the latest context from the daily memory and let's resume work on the auth module.'
Tips & Limitations
- Tips: Leverage the
memory_searchutility to quickly query across files. Treat the 'Permanent Memory' as your source of truth for identity and technical constraints. Perform a weekly review to purge noise, keeping the memory pool focused and relevant. - Limitations: This is a filesystem-based system, meaning it lacks the semantic fuzzy-matching power of a vector database out of the box. Users should avoid storing overly sensitive credentials or large, non-essential binary logs that could clutter the directory. While it excels at structure and context, extremely massive data repositories may require a transition to a hybrid vector-search model for optimal performance.
Metadata
Not sure this is the right skill?
Describe what you want to build — we'll match you to the best skill from 16,000+ options.
Find the right skillPaste this into your clawhub.json to enable this plugin.
{
"plugins": {
"official-daoistbro-memory-system": {
"enabled": true,
"auto_update": true
}
}
}Tags(AI)
Flags: file-write, file-read