← 返回 Skills 市场
Project Sharing System
作者
Clement Gu
· GitHub ↗
· v1.0.0
· MIT-0
73
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install project-sharing-system
功能描述
项目共享系统 — 多 Agent 项目状态协作与自动发现
安全使用建议
This package mostly does what its name says (CLI, snapshots, backups), but review these before installing:
- Paths & env: scripts honor CLAWHUB_WORKDIR and AGENT_ID but the skill metadata declares no env vars. Confirm where projects_status.json and PROJECT_STATUS.md will be stored (user workspace vs. skill directory).
- File writes: the tool will write/overwrite projects_status.json, PROJECT_STATUS.md, create backups and append to a 'memory' folder. Make backups of any existing project files first.
- Ambiguity: update_project_status.js uses __dirname.. to locate projects_status.json which could cause it to modify files bundled with the skill instead of your workspace. Search/modify these paths to point to a safe workspace before using.
- Run in a sandbox first: execute project sync / discover_projects.sh in a controlled environment (or with a test CLAWHUB_WORKDIR) to observe where files are created.
- Minor bugs: there are small code inconsistencies (e.g., property names used in log entries) — expect noisy logs and test behavior.
If you accept these caveats, install in a safe workspace and/or adjust the WORKSPACE/CLAWHUB_WORKDIR to a directory you control.
功能分析
Type: OpenClaw Skill
Name: project-sharing-system
Version: 1.0.0
The bundle implements a project management and status-sharing system designed for multi-agent environments. It provides CLI tools (project.js, project.sh), status discovery scripts (discover_projects.sh), and automated backup logic (auto_backup.sh) to maintain a centralized JSON database and Markdown reports. The code uses standard file I/O and local script execution to synchronize project states, and no indicators of data exfiltration, unauthorized network access, or malicious prompt injection were found.
能力评估
Purpose & Capability
Name/description (project-sharing, multi-agent project status) matches the included CLI, snapshot and backup scripts. The requested capabilities (list, show, add, update, sync, snapshot) align with the code. However some code paths reference different workspace locations (defaults to $HOME/.openclaw/workspace vs. __dirname/..), which is inconsistent with a single clear runtime workspace.
Instruction Scope
SKILL.md instructs Agents to read snapshots and use CLI; the scripts do exactly that (generate snapshots, print agent-parsable lines). They read/write projects_status.json, PROJECT_STATUS.md, create backups and memory files. No network endpoints or secret exfiltration are present. But the instructions grant the skill broad discretion to read and write multiple files in the user's CLAWHUB_WORKDIR or $HOME/.openclaw/workspace and to append to a 'memory' directory—actions not explicitly declared in SKILL.md.
Install Mechanism
No install spec — instruction-only with shipped scripts. That limits external code download risk. All code is bundled with the skill (no remote fetches or installers).
Credentials
SKILL metadata declares no required env vars, but the scripts use environment variables (CLAWHUB_WORKDIR, AGENT_ID). They default to $HOME/.openclaw/workspace or derive WORKSPACE from __dirname, so environment affects behavior. The skill also writes into a 'memory' directory and may append daily memory files. These env usages and file writes are reasonable for a project tool but are not declared and could cause the skill to read/write unexpected locations depending on environment.
Persistence & Privilege
always:false (normal). The skill writes/updates data files (projects_status.json, PROJECT_STATUS.md), creates backups and memory logs. One inconsistency: update_project_status.js writes to PROJECTS_JSON_PATH defined as path.join(__dirname, '..', 'projects_status.json'), which may modify the installed package's bundled JSON (i.e., files inside the skill directory) rather than a user workspace. That behavior is ambiguous and may result in the skill modifying its own shipped files or unexpected locations; it's not an overt privilege escalation but should be understood.
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 project-sharing-system - 安装完成后，直接呼叫该 Skill 的名称或使用
/project-sharing-system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
v1.0.0
初始发布：多Agent项目状态共享、CLI管理、自动备份、Agent自动发现
元数据
常见问题
Project Sharing System 是什么？
项目共享系统 — 多 Agent 项目状态协作与自动发现. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 73 次。
如何安装 Project Sharing System？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install project-sharing-system」即可一键安装，无需额外配置。
Project Sharing System 是免费的吗？
是的，Project Sharing System 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Project Sharing System 支持哪些平台？
Project Sharing System 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 Project Sharing System？
由 Clement Gu（@clementgu）开发并维护，当前版本 v1.0.0。
推荐 Skills