← 返回 Skills 市场
71
总下载
0
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install core-capabilities-v2
功能描述
工作助手核心能力集成包 - 包含 Obsidian/Git 同步、记忆数据库、自然语言查询工具、监控页面等完整能力。提供 memory_query_agent.py 工具和完整文档。
使用说明 (SKILL.md)
Core Capabilities v2 - 工作助手核心能力完整包
本技能包整合了工作助手的所有核心能力，提供完整的工具集、查询能力和使用指导。
📦 包含工具
- memory_query_agent.py - 自然语言记忆查询工具
- monitor_server.py - Web 监控服务器
- setup_cron.sh - Cron 配置脚本
🎯 核心能力
- 🧠 Obsidian 和 Git 同步 - 完整的文件管理和版本控制
- 📊 记忆数据库 - SQLite 存储 + 自动同步（24 条记录）
- 🔍 自然语言查询 - 中文智能查询
- 📈 监控页面 - Web 实时监控
🚀 使用
# 查询记忆
python3 memory_query_agent.py "最近的记录"
# 交互模式
python3 memory_query_agent.py -i
# 查看状态
python3 memory_query_agent.py --sync-status
📊 状态
- 数据库：24 条记录
- 同步：每 30 分钟
- 监控：8003 端口
💡 示例
用户：我们有哪些能力？
助手：我们有四大核心能力...
版本: 2.0.0
创建: 2026-04-12
安全使用建议
This package implements the advertised features but contains risky defaults and inconsistencies. Before installing or running: (1) Inspect memory_query_agent.py, monitor_server.py and setup_cron.sh locally to confirm where data will be read/written. (2) Note monitor_server.py uses WORKSPACE = /home/awu/.openclaw/workspace (hardcoded) — change this to a safe directory or the script's directory if you plan to run it. (3) The monitor binds to all network interfaces and sets CORS to '*', and /api/memories returns full memory content — run it only on localhost or behind firewall, or change the server to bind to 127.0.0.1. (4) The server executes cron_monitor.py in that workspace via os.system; ensure that file is trusted before allowing it to run. (5) Do not run setup_cron.sh unless you want a persistent cron job; prefer running the agent manually or create a containerized/sandboxed environment for testing. (6) Consider backing up and scanning the memory.db for sensitive data before exposing it externally. If you want, I can list the exact lines that show the hardcoded path, the os.system call, and the HTTP/CORS exposure to help you modify the code safely.
功能分析
Type: OpenClaw Skill
Name: core-capabilities-v2
Version: 2.0.1
The skill bundle provides legitimate-seeming memory management and monitoring functionality but employs several high-risk patterns. Key indicators include the use of os.system() in monitor_server.py to execute shell commands, the modification of the user's crontab for persistence via setup_cron.sh, and the opening of a local network port (8000/8003) to serve database content. Additionally, monitor_server.py contains a hardcoded absolute path (/home/awu/.openclaw/workspace), which is a security anti-pattern and suggests the code may not be safely portable.
能力评估
Purpose & Capability
The package claims Obsidian/Git sync, memory DB, query tool and monitoring and does include a memory agent, monitor server and cron script. However the pieces are inconsistent: memory_query_agent.py uses a memory directory and memory.db colocated with the script, while monitor_server.py expects files under a hardcoded WORKSPACE = /home/awu/.openclaw/workspace. That mismatch suggests sloppy design or an assumption about the runtime environment that isn't declared in the skill metadata.
Instruction Scope
Runtime behavior includes reading local markdown files and an SQLite database and returning full 'content' fields via an HTTP API. monitor_server.py binds a TCPServer to "" (all interfaces) and sets Access-Control-Allow-Origin: *, making potentially sensitive memory content reachable from the network. The server also runs os.system('cd WORKSPACE && python3 cron_monitor.py ...'), invoking an arbitrary script in that workspace. These actions go beyond a simple local query tool and can expose or execute unverified code.
Install Mechanism
There is no formal install spec (lowest risk from registry perspective). However the package includes setup_cron.sh which, if run, will modify the user's crontab to run the agent every 30 minutes. That is a local persistence mechanism and should be considered a privileged operation that the user must consciously approve.
Credentials
The skill declares no required environment or credentials, but monitor_server.py hardcodes access to /home/awu/.openclaw/workspace and expects cron_monitor.py and memory.db there. The code will read workspace files (cron_status.json, memory.db, html) and may execute workspace scripts. These implicit accesses to a user workspace are not declared and could expose unrelated/secret files.
Persistence & Privilege
The skill does not set always:true, but it ships with a script that, when executed by the user, will add a cron job to persistently run the agent. The monitoring server also opens a network service (binds to all interfaces by default). Combined, these allow ongoing background access to local memories and network exposure if the user accepts the setup script.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install core-capabilities-v2 - 安装完成后，直接呼叫该 Skill 的名称或使用
/core-capabilities-v2触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v2.0.1
- Added three new files: memory_query_agent.py, monitor_server.py, and setup_cron.sh.
- No significant user-facing changes in documentation or SKILL.md content.
- Prepares the skill for enhanced memory query, monitoring, and automation features.
v2.0.0
Core Capabilities v2.0.0 – 集成 Obsidian/Git 同步、记忆数据库、自然语言查询、监控等核心功能：
- 新增 memory_query_agent.py，自然语言记忆查询工具，支持交互与多种状态操作
- 集成 Obsidian 与 Git 全面同步，提供版本控制与管理
- 内置记忆数据库（SQLite），自动同步，每 30 分钟更新
- 提供 Web 监控页面（默认端口 8003）
- 完整文档和使用示例，便于用户快速上手
元数据
常见问题
Core Capabilities V2 是什么？
工作助手核心能力集成包 - 包含 Obsidian/Git 同步、记忆数据库、自然语言查询工具、监控页面等完整能力。提供 memory_query_agent.py 工具和完整文档。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 71 次。
如何安装 Core Capabilities V2？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install core-capabilities-v2」即可一键安装，无需额外配置。
Core Capabilities V2 是免费的吗？
是的，Core Capabilities V2 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Core Capabilities V2 支持哪些平台？
Core Capabilities V2 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Core Capabilities V2？
由 awublack（@awublack）开发并维护，当前版本 v2.0.1。
推荐 Skills