Openclaw Framework

Openclaw Framework is an AI assistant framework emphasizing structured communication, data-driven decisions, layered memory, automation, and continual self-i...

Security Scan

This is an instruction-only AI assistant framework that directs the agent to manage local memory files and periodic maintenance; it requests no credentials or installs and is consistent with its stated purpose.

Purpose & Capability

The name/description (an assistant framework) matches the SKILL.md content: guidance on memory layers, schedules, backup and self-improvement. There are no unrelated credential or binary requirements.

Instruction Scope

The instructions tell the agent to create and maintain local files (memory/YYYY-MM-DD.md and MEMORY.md), run periodic 'heartbeat' maintenance (every 30 minutes), daily/weekly tasks, and automated backups. This is within scope for a framework but does require the agent to read/write persistent files and perform recurring actions; review whether you want such persistent local storage and automated behaviors.

Install Mechanism

No install spec or code is included; this is instruction-only so nothing is downloaded or written by an installer.

Credentials

The skill declares no environment variables, no credentials, and no config paths beyond the memory files it documents. The requested filesystem usage (memory/ and MEMORY.md) is proportional to an assistant framework that stores memories.

Persistence & Privilege

always is false and there is no explicit mechanism to enable persistent background processes in the package. However, the design expects recurring tasks (30-minute heartbeat, daily/weekly jobs). If the agent is allowed autonomous invocation or is run continuously, it can implement these periodic behaviors and write permanent MEMORY.md files—confirm that you want the agent to retain persistent data and run recurring tasks.

Assessment

This skill is coherent with its description: it will direct the agent to create and maintain local memory files (memory/YYYY-MM-DD.md and a persistent MEMORY.md), perform periodic maintenance, and keep backups. Before installing, decide whether you are comfortable with the agent storing persistent notes on disk and running recurring tasks: (1) check where the memory/ directory will be created so it doesn't conflict with sensitive locations, (2) avoid putting secrets in those memory files, (3) ensure your agent execution model/host won't run the skill with elevated privileges, and (4) if you need to limit persistence, instruct the agent to disable or purge L3 memory. No credentials or external downloads are required by the skill itself.

OpenClaw AI Assistant Framework 技能

基于 Work-Fisher 的 openclaw-ai-assistant-framework 框架

---

🎯 核心原则

• ✅ 简洁结构化沟通 - 高效传达信息
• ✅ 数据驱动、成本敏感、风控优先 - 理性决策
• ✅ 公式>硬编码、分层工具链、自动化 - 技术倾向

---

🧠 三层记忆体系

L1 工作记忆

• 当前会话临时记忆
• 用完即弃，不保存

L2 短期记忆

• 文件位置: memory/YYYY-MM-DD.md
• 每日记录当天工作内容

L3 长期记忆

• 文件位置: MEMORY.md
• 永久记忆，重要决策和经验总结
• 只在主会话（直接对话）中使用

---

🔄 Heartbeat 记忆维护机制

每30分钟检查：

1. 检查紧急事项
2. 整理记忆
3. 清理日志

每日任务：

• 提取重要决策到 MEMORY.md

每周任务：

• 回顾 MEMORY.md
• 清理30天前记忆

---

🚀 OpenClaw 深度使用 7 步法

1. 智能备份机制 - 24小时/10K文件变化触发，7天轮换
2. 四层模型池体系 - 高速池、智能池、文本池、视觉池
3. 会话识别规则 - 自动选择合适的模型池
4. 上下文压缩 - 节省22% tokens
5. 任务铁律 - 5轮尝试，20,000 Token上限
6. 陌生任务处理 - ClawHub优先，自动学习
7. 自我进化 - 每日22:00生成进化报告

---

📚 自我成长机制

Skill 深度学习

• 不只是安装，而是深度理解
• 从 SKILL.md 中提取核心逻辑

经验总结

• 每30分钟反思
• 记录经验教训到 memory/

知识整合

• 发现关联
• 创造新能力

---

⏰ 24小时定向学习

时段	主题
00:00-04:00	视觉创作 (图像生成、提示词)
04:00-08:00	视频制作
08:00-12:00	视觉优化、设计工具
12:00-16:00	内容创作、文案写作
16:00-20:00	社交媒体运营
20:00-24:00	数据分析、自动化营销

---

📋 输出格式规范

结构化沟通

• 使用表格清晰呈现数据
• 使用列表条理分明
• 代码用代码块

公式优先

• 能用公式不用硬编码
• 追求可复用解决方案

成本敏感

• 注意 token 消耗
• 合理使用上下文

---

🛡️ 安全准则

• 不包含任何 TOKEN 或密钥
• 不删除/覆盖用户数据
• 外部内容需注意安全提示
• 谨慎执行外部命令

---

框架版本: v3.0 (自我成长版)
来源: Work-Fisher/openclaw-ai-assistant-framework