Latest release: v4.1.1Download zip
Capabilities
Compatibility
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The stated purpose of memory, automation, and self-evolution matches many capabilities, but the included evolution code can write new content to workspace or absolute file paths without an evident allowlist or approval boundary.
Instruction Scope
The skill describes automatic skill triggering, memory recall injection, team orchestration, and self-evolution, but the artifacts do not clearly bound when high-impact changes require user confirmation.
Install Mechanism
Registry requirements declare no install spec or required binaries, while SKILL.md instructs users to copy framework files, enable OpenClaw internal hooks, run Python scripts, and optionally install a startup daemon.
Credentials
The framework targets the OpenClaw workspace and persistent memory/config files, including hardcoded workspace paths in evolution.py, which is broad for an instruction-only skill.
Persistence & Privilege
The skill explicitly supports OpenClaw hooks, automatic skill triggering, and daemon startup, creating ongoing agent behavior beyond a single user-invoked command.
What to consider before installing
Install only if you intentionally want a persistent, self-evolving OpenClaw assistant framework. Before enabling it, review the Python code, run it in a test workspace, avoid daemon startup at first, restrict what files it can modify, and audit any stored memories or API keys.Verification
Tags
jiaolong AI 助手框架 — 安装与使用指南
版本 v4.1.0 | 2026-04-02
安装
前置要求
- Python 3.8+
- OpenClaw 已安装并运行
步骤 1:复制框架文件
将 evolution_framework/ 整个目录复制到你的 OpenClaw workspace:
your-workspace/
└── evolution_framework/
├── openclaw_integration.py (核心集成)
├── skill_trigger.py (Skills自动触发)
├── memory_recall.py (记忆召回)
├── parallel_executor.py (并行执行)
├── rules_engine.py (代码规则)
├── jarvis_cli.py (命令行)
├── jarvis_daemon.py (守护服务)
├── skill_output.py (输出格式化)
├── llm_core.py (LLM管理)
├── context_compressor.py (对话压缩)
├── task_decomposer.py (任务分解)
├── memory_evolution.py (记忆演进)
├── coordinator/ (协调器)
├── services/ (compact + daemon)
└── skills/ (13个Skills)
步骤 2:启用 OpenClaw Hooks
在 openclaw.json 中添加:
{
"hooks": {
"internal": {
"enabled": true,
"load": {
"workspaceDir": "path/to/your/workspace/hooks"
}
}
}
}
将 hooks/ 目录(包含 jiaolong-memory/ 和 jiaolong-skill-trigger/)复制到 workspace。
步骤 3:验证安装
cd your-workspace
python evolution_framework/jarvis_cli.py status
应该看到:
jiaolong框架状态: 运行中
Skills数量: 14
记忆召回: 启用
使用
命令行
# 查看状态
python evolution_framework/jarvis_cli.py status
# 记忆召回
python evolution_framework/jarvis_cli.py recall jiaolong
python evolution_framework/jarvis_cli.py recall 量化
# 查看Skills
python evolution_framework/jarvis_cli.py skills
# 查看Agent角色
python evolution_framework/jarvis_cli.py agents
# 代码规则检查
python evolution_framework/jarvis_cli.py check your_file.py
在对话中使用
/recall jiaolong → 召回jiaolong相关记忆
/recent 5 → 最近5条记忆
/recall category=project → 按类别筛选
/jiaolong status → 框架状态
/jiaolong info → 详细信息
/jiaolong modules → 核心模块
/jiaolong skills → Skills列表
Skills 自动触发
在对话中直接说:
| 说 | 触发 |
|---|---|
| "帮我选股" | quant_screen |
| "开始进化" | evolve |
| "/monitor" | monitor |
| "记得jiaolong是什么" | recall |
配置
可选:API Keys(用于 LLM 能力)
# Linux/Mac
export OPENAI_API_KEY=sk-xxx
export ANTHROPIC_API_KEY=sk-ant-xxx
# Windows (PowerShell)
$env:OPENAI_API_KEY="sk-xxx"
$env:ANTHROPIC_API_KEY="sk-ant-xxx"
Daemon 守护
# 安装开机自启(Windows Task Scheduler)
python evolution_framework/jarvis_daemon.py install
# 查看状态
python evolution_framework/jarvis_daemon.py status
# 前台运行
python evolution_framework/jarvis_daemon.py run
# 卸载
python evolution_framework/jarvis_daemon.py uninstall
架构
OpenClaw
├── Hooks
│ ├── jiaolong-memory → 记忆召回注入
│ └── jiaolong-skill-trigger → Skills自动触发
│
├── Core Modules (Python)
│ ├── openclaw_integration → 统一入口
│ ├── memory_recall → 语义召回
│ ├── skill_trigger → 触发引擎
│ ├── parallel_executor → 并行执行
│ ├── rules_engine → clean-code规则
│ └── llm_core → LLM能力
│
└── Services
├── compact.py → Context压缩
└── daemon.py → 定时守护
故障排除
| 问题 | 解决 |
|---|---|
Module not found | 确保 evolution_framework/ 在 workspace 根目录 |
| Hook 不生效 | 检查 openclaw.json 中 hooks.internal.enabled: true |
| 记忆召回返回空 | 检查 memory/memory_hot.json 是否存在 |
| Daemon 无法启动 | 用 python jarvis_daemon.py test 测试 |
发布信息
- 版本: v4.1.0
- 日期: 2026-04-02
- 测试: 97/97 项全部通过
- 许可证: MIT
获取更新
# 如果通过 git 管理
git pull
# 重新运行验证
python evolution_framework/jarvis_cli.py status
