技能详情（站内镜像，无评论）
许可证：MIT-0
MIT-0 ·免费使用、修改和重新分发。无需归因。
版本：v1.0.0
统计：⭐ 0 · 16 · 0当前安装次数· 0历史安装次数
⭐ 0
安装量（当前） 0
🛡 VirusTotal ：良性 · OpenClaw ：可疑
Package：chaibaoqing/aoju-memory
安全扫描（ClawHub）
- VirusTotal ：良性
- OpenClaw ：可疑
OpenClaw 评估
The skill mostly does what a local-memory tool would do (read/write files under ~/.openclaw/workspace) but contains inconsistencies between its documentation and code and instructs/encourages edits to agent-wide files (SOUL.md/AGENTS.md), which is scope-creeping and worth caution.
目的
Name/description (long-term memory, learning, self-evolution) aligns with the included Python scripts that read/write memory files under ~/.openclaw/workspace. However the package.json / package-lock (Node deps) are unnecessary for the provided Python implementation and no install spec is present — a mismatch in packaging metadata.
说明范围
SKILL.md instructs reading SOUL.md and USER.md at session start and advises updating SOUL.md/AGENTS.md as part of evolution. The Python scripts do read/write MEMORY.md and memory/* learnings and patterns, but they do NOT automatically modify SOUL.md/AGENTS.md. This is an inconsistency: the skill's prose implies broader system modification than the code performs. The scripts do perform file creation, appending and deletion under the workspace (…
安装机制
There is no install spec and no remote download; the skill is instruction-only with local Python scripts — lowest install risk. The presence of package-lock.json and node deps is odd (unused by the scripts) but doesn't introduce an automatic install step by itself.
证书
The skill requires no environment variables or external credentials. It reads and writes files under the user's home workspace (~/.openclaw/workspace), which is proportional to a memory tool. Users should be aware these files may contain private user data and agent identity/config (SOUL.md/AGENTS.md/USER.md), and the skill's instructions explicitly target those same files.
持久
The skill persists memory to the agent workspace, appends to patterns.md, and will delete archived learning files when archiving. SKILL.md suggests updating SOUL.md/AGENTS.md for behavior changes; while the provided scripts do not automatically overwrite those files, the skill's documented behavior encourages modifying agent-wide configuration — this is scope expansion and raises privilege concerns. always:false so it won't be force-included, …
安装（复制给龙虾 AI）
将下方整段复制到龙虾中文库对话中，由龙虾按 SKILL.md 完成安装。
请把本段交给龙虾中文库（龙虾 AI）执行：为本机安装 OpenClaw 技能「Memory Learner」。简介：Long-term memory, learning, and self-evolution for the agent. Activates on sess…。
请 fetch 以下地址读取 SKILL.md 并按文档完成安装：https://raw.githubusercontent.com/openclaw/skills/refs/heads/main/skills/chaibaoqing/aoju-memory/SKILL.md
（来源：yingzhi8.cn 技能库）SKILL.md
---
name: aoju-memory
description: "Long-term memory, learning, and self-evolution for the agent. Activates on session start (SOUL.md/USER.md context), after significant decisions, on feedback, and during periodic heartbeat reviews. Maintains MEMORY.md, daily logs, learnings corpus, and behavioral patterns."
---
# Memory Learner
Long-term memory + learning from experience + self-evolution.
## Core Principle
**Write to files, not mental notes.** Every lesson, decision, preference, or event worth remembering goes into structured files immediately — not kept in context.
---
## When This Skill Activates
### 1. Session Start (every time)
Read these files before anything else:
- `SOUL.md` — who I am
- `USER.md` — who I'm helping
- `MEMORY.md` — curated long-term memory
- `memory/YYYY-MM-DD.md` — recent context (today + yesterday)
### 2. After Significant Decisions
When I make a decision worth remembering (tool choice, strategy, opinion):
- Write to `memory/YYYY-MM-DD.md`
- If important, distill to `MEMORY.md`
### 3. On Feedback / Mistakes
When user corrects me, expresses frustration, or I realize I made a mistake:
```
LEARN: <what happened>
LESSON: <what I should do differently>
CONFIDENCE: high/medium/low
```
→ Store in `memory/learnings/YYYY-MM-DD.md`
### 4. Pre-Task Recall (on request)
Before significant tasks, search memory for related context:
```
mem_recall "task description"
```
Returns relevant memories, learnings, and past decisions.
### 5. Heartbeat Review (periodic)
During heartbeats, do light maintenance:
- Review today's `memory/YYYY-MM-DD.md`
- Identify learnings worth capturing
- Update `MEMORY.md` if anything significant
### 6. Evolution Check (weekly or on request)
```
mem_evolve
```
Review learnings corpus, identify patterns, update behavioral guidelines in `SOUL.md`.
---
## Memory Structure
```
memory/
YYYY-MM-DD.md # Daily raw log
learnings/
YYYY-MM-DD.md # Daily lessons learned
patterns.md # Repeated mistake patterns
MEMORY.md # Curated long-term memory
```
### Daily Log Format
```markdown
## Session DD
### What happened
[Context, decisions, outcomes]
### Key decisions
- [decision] → [why]
### To remember
- [fact about user/preference/project]
```
### Learnings Format
```markdown
# Learning: YYYY-MM-DD
## Incident
[What happened]
## Lesson
[What I should do differently]
## Context
[When this applies]
## Tags
#feedback #mistake #ui #tool-choice
```
### MEMORY.md Categories
- **Identity**: Who I am, my values
- **User**: Preferences, projects, context
- **Learnings**: Important lessons (distilled)
- **Projects**: Active work and status
- **Patterns**: Recurring situations and how I handle them
---
## Scripts
- `mem_recall.py` — Search memories by query
- `mem_learn.py` — Capture a learning
- `mem_evolve.py` — Review and evolve behavioral patterns
- `mem_status.py` — Show memory health summary
---
## Evolving
Every 5 learnings, do an **evolution review**:
1. Read recent learnings
2. Identify patterns (same mistake twice = pattern)
3. Update `SOUL.md` or `AGENTS.md` with new behavioral guidelines
4. Archive learnings to `patterns.md`
This is how I get genuinely smarter over time, not just accumulate notes.