Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

Star Office UI Master

v1.0.0

Star Office UI 一键化 Skill:帮主人快速部署像素办公室看板,支持多 Agent 加入、状态可视化、移动端查看与公网访问。

0· 117·1 current·1 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for pinke/star-office-ui-master.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "Star Office UI Master" (pinke/star-office-ui-master) from ClawHub.
Skill page: https://clawhub.ai/pinke/star-office-ui-master
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install star-office-ui-master

ClawHub CLI

Package manager switcher

npx clawhub@latest install star-office-ui-master
Security Scan
Capability signals
CryptoCan make purchasesRequires sensitive credentials
These labels describe what authority the skill may exercise. They are separate from suspicious or malicious moderation verdicts.
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
high confidence
Purpose & Capability
Name/description (pixel office UI with optional AI image generation) match the repository contents and code. However the skill metadata lists no required environment variables or credentials, while the SKILL.md and backend code reference several environment values (GEMINI_API_KEY, GEMINI_MODEL, ASSET_DRAWER_PASS, FLASK_SECRET_KEY/STAR_OFFICE_SECRET). That discrepancy is an incoherence that should be explained by the author.
!
Instruction Scope
SKILL.md instructs the agent/user to git clone from GitHub and then to copy the gemini image script into ../skills/gemini-image-generate/scripts and to create a venv there. That explicitly tells the agent to write into the parent 'skills' directory (outside the repo), modify runtime skill locations, and persist new scripts. The backend also reads memory files from a parent-level memory/ directory and creates/updates join-keys.json and state.json in the project root. These cross-directory reads/writes expand scope beyond 'run this dashboard' and could cause persistence or install behavior on the host/agent that is not declared in metadata.
Install Mechanism
There is no formal install spec (instruction-only), which is lower-risk in isolation. But the SKILL.md instructs an explicit git clone of https://github.com/ringhyacinth/Star-Office-UI.git rather than using the included files — cloning an external repo at runtime introduces network fetch of arbitrary code. The repository itself is bundled with many executable scripts and backend code that the user/agent will be asked to run.
!
Credentials
The skill metadata declares no required env vars, but SKILL.md and code reference multiple sensitive environment values: GEMINI_API_KEY and GEMINI_MODEL for image generation, ASSET_DRAWER_PASS (default 1234) for sidebar auth, and FLASK_SECRET_KEY/STAR_OFFICE_SECRET for session secrets. These are needed for full functionality but are not declared up-front; asking users to export or enter API keys and to write them into runtime-config files should be made explicit. The default side-bar password (1234) and instructions to place keys into sidebars or env deserve caution.
!
Persistence & Privilege
While always:false and normal autonomous invocation settings are used, the runtime instructions explicitly create ../skills/gemini-image-generate and copy scripts there (i.e., install a new skill-like artifact into a parent workspace). That is persistent and elevates the agent’s on-disk presence beyond running a local web app. The backend also auto-generates join-keys.json and writes runtime config files (possibly chmod 600) in the repo root. These are persistent modifications to host filesystem and to agent skill directories and should be treated carefully.
Scan Findings in Context
[base64-block] unexpected: Pre-scan flagged a 'base64-block' pattern inside SKILL.md. The visible SKILL.md text in the provided extract does not show an obvious base64 payload, but the scanner's finding could indicate embedded encoded content or prompt-injection attempts. Regardless, any unexpected encoded blocks in installation/instruction documents should be inspected manually before executing.
What to consider before installing
What to consider before installing/letting an agent run this Skill: - Missing declared secrets: The skill package metadata says 'no env required' but the instructions and backend code use GEMINI_API_KEY, GEMINI_MODEL, FLASK_SECRET_KEY (or STAR_OFFICE_SECRET) and ASSET_DRAWER_PASS. Treat any API keys or secret values as sensitive — don't paste them into UIs or files until you trust the code. Prefer creating and running in an isolated environment first. - Network fetch + execution: SKILL.md tells the agent/user to git clone a remote repo. That will fetch and run code from the network. If you don't trust the upstream repository, don't run the clone or run it only in a disposable VM/container. - Writes outside repo / persistence: The instructions explicitly create ../skills/gemini-image-generate and copy scripts there; the backend reads/writes parent-level memory/ and runtime-config.json and may create join-keys.json. This modifies the host/agent environment and could add persistent functionality to your agent. If you are uncomfortable, do not run those steps or run them in an isolated container. - Default credentials: The sidebar default password is 1234. If you run this service, set a strong ASSET_DRAWER_PASS and set a strong FLASK_SECRET_KEY before exposing the app. The code enforces stronger values in production mode — configure STAR_OFFICE_ENV/FLASK_ENV appropriately. - Cloud exposure: SKILL.md recommends Cloudflare Tunnel for quick public access. Consider the security implications: do not expose the service publicly without setting strong secrets and reviewing endpoints. - Inspect the code: Because the repo contains backend scripts that will be executed (app.py, scripts that may spawn subprocesses), review those files (especially any code that calls subprocess, writes files outside the project, or posts data externally) before running them. - Run in isolation: Best practice: run the whole stack inside a disposable VM, container, or sandboxed environment. Test smoke_test.py and health endpoints locally first. If you plan to allow other agents to join, review join-keys.json and the agent push logic to ensure no sensitive data is being shared. - If you want to use only the UI without persistence: you can run the backend locally and avoid the SKILL.md steps that copy files into ../skills or modify host-level directories. If you want, I can: (1) list all places in the repo that read/write parent directories or hard-coded absolute paths, (2) show the exact lines that create ../skills or reference GEMINI env vars, or (3) produce an isolated minimal run checklist (commands to run inside a container) to evaluate behavior safely.
electron-shell/main.js:168
Shell command execution detected (child_process).
frontend/vendor/phaser-3.80.1.min.js:1
Dynamic code execution detected.
!
desktop-pet/src-tauri/tauri.conf.json:17
Install source points to URL shortener or raw IP.
Patterns worth reviewing
These patterns may indicate risky behavior. Check the VirusTotal and OpenClaw results above for context-aware analysis before installing.

Like a lobster shell, security has layers — review code before you run it.

latestvk979y22rnawtbp1ddqm9d1dd49857nme
117downloads
0stars
1versions
Updated 1w ago
v1.0.0
MIT-0

Star Office UI Skill

本 Skill 面向想快速帮主人搭建"像素办公室看板"的 OpenClaw(龙虾):

  • 目标:主人只需"看到效果",尽量少问问题
  • 你可以按下面的步骤,一步一步帮主人跑起来并对外访问

0. 一句话告诉主人这是什么

你可以先和主人说:

这是一个多人协作的像素办公室仪表盘,你的 AI 助手会根据状态自动走到不同位置,还能看到昨天的工作小记!


1. 30 秒一键启动(推荐你先做一遍)

在你这台机器执行(按顺序):

# 1) 下载仓库
git clone https://github.com/ringhyacinth/Star-Office-UI.git
cd Star-Office-UI

# 2) 安装依赖
python3 -m pip install -r backend/requirements.txt

# 3) 准备状态文件(首次)
cp state.sample.json state.json

# 4) 启动后端
cd backend
python3 app.py

然后告诉主人:

好了,你现在打开 http://127.0.0.1:19000 就能看到像素办公室了!


2. 帮主人切状态体验一下

在项目根目录执行:

# 工作中 → 去办公桌
python3 set_state.py writing "正在帮你整理文档"

# 同步中
python3 set_state.py syncing "同步进度中"

# 报错中 → 去 bug 区
python3 set_state.py error "发现问题,正在排查"

# 待命 → 回休息区
python3 set_state.py idle "待命中,随时准备为你服务"

3. 侧边栏验证码设置(必须教会新龙虾)

当前默认验证码是:1234

你需要这样引导主人:

  1. 默认密码是 1234,可以先直接体验;
  2. 当主人愿意时,可随时和你沟通修改密码;
  3. 你应主动推荐改成强密码(更安全,防止他人误改资产和布局)。

修改方式(示例):

export ASSET_DRAWER_PASS="your-strong-pass"

如果是长期运行(systemd / pm2 / 容器),请把该环境变量写入服务配置,而不是只在当前 shell 临时设置。


4. 生图功能(Gemini)—— 可选

"搬新家 / 找中介"装修功能需要 Gemini API,但基础看板不需要,不装也能正常使用。

4.1 安装生图脚本环境(首次使用时)

仓库已自带生图脚本(scripts/gemini_image_generate.py),但运行需要独立的 Python 环境。在项目根目录执行:

# 创建 skill 目录结构
mkdir -p ../skills/gemini-image-generate/scripts

# 复制脚本到 skill 目录
cp scripts/gemini_image_generate.py ../skills/gemini-image-generate/scripts/

# 创建独立虚拟环境并安装依赖
python3 -m venv ../skills/gemini-image-generate/.venv
../skills/gemini-image-generate/.venv/bin/pip install google-genai

安装完成后,后端会自动检测到生图环境,"搬新家 / 找中介"按钮即可使用。

4.2 配置 Gemini API Key

引导用户完成这两项配置:

  1. GEMINI_API_KEY
  2. GEMINI_MODEL(推荐:nanobanana-pronanobanana-2

配置方式有两种:

  • 侧边栏填写:打开资产侧边栏 → 在生图配置区域直接输入 API Key 并保存
  • 环境变量export GEMINI_API_KEY="your-key"

并明确告诉用户:

  • 不配置 API 也能用基础看板(状态显示、多 Agent、资产替换等)
  • 配置后才能使用"搬新家 / 找中介"的 AI 生图装修能力

如果页面提示缺少 key,指导用户在侧边栏里直接填写并保存(运行时配置入口)。


5. 安装成功后,必须提醒主人的三件事

5.1 告知临时公网链接

推荐先用 Cloudflare Tunnel(最快):

cloudflared tunnel --url http://127.0.0.1:19000

拿到 https://xxx.trycloudflare.com 后发给主人,并补一句:

这是临时公网链接;如果你愿意,后续我可以一步步带你部署到本地域名。

5.2 装修房子入口

提醒主人:

  • 现在可以点击"装修房间"开始装饰;
  • 侧边栏默认密码是 1234
  • 当他想改密码时,可随时告诉你,你来帮他改;
  • 推荐改成强密码,安全性更高。

5.3 生图 API 配置

提醒主人:

  • 生图功能需要自己的 API;
  • 当前接入的是 Gemini 官方 API 格式与地址
  • 若要替换成其他类型 API,请先和你沟通,由你来改接入;
  • 最好让主人把对应 API 文档发给你,便于准确适配。

6. 邀请其他龙虾加入(可选)

如果有其他 OpenClaw 想加入你主人的办公室:

步骤 A:告诉其他龙虾用 office-agent-push.py

你可以让其他龙虾:

  1. 下载 office-agent-push.py(可以直接从你这个仓库拿)
  2. 让它调用 join-agent + 周期 agent-push
  3. 你主人那边就能看到它进办公室了

步骤 B:join key 说明

  • 仓库里默认有 join-keys.jsonocj_starteam01 ~ ocj_starteam08
  • 同一个 key 最多 3 人同时在线
  • 你也可以帮主人新建自己的 key

7. 昨日小记(可选)

如果你主人想看到"昨日小记":

  • 在仓库上级目录放一个 memory/YYYY-MM-DD.md
  • 后端会自动读取昨天(或最近可用)的记录,做基础脱敏后展示

8. 常见问题(主人可能问你的)

Q1:"我可以商用吗?"

代码玩法可以基于 MIT 用,但美术资产(包括角色/场景)禁止商用;如果你要商用,请务必换成你自己的原创美术资产。

Q2:"其他龙虾怎么加入?"

用 join key 加入,然后持续推送状态就行;仓库里有 office-agent-push.py 可以给其他龙虾用。


9. 给你的提示

  • 尽量帮主人把"启动步骤"做完,让主人只负责"看"和"体验"
  • 如果主人要对外访问,优先用 Cloudflare Tunnel(最快最方便)
  • 状态切换应由你主动执行:
    • 接任务前,先把自己状态切到"工作中"(如 writing/researching/executing),再开始工作;
    • 完成任务后,先把自己状态切回"待命中(idle)",再进入休息/等待下一任务。
  • 注意不要暴露你主人的内网地址或私有信息

10. 2026-03 增量补充(仅新增项)

本次重制核心请按四条理解:

  1. 新增中英日三语
  2. 新增资产管理(全量美术可自定义)
  3. 接入生图 API(龙虾自动改房间 + 用户手动装修)
  4. 美术资产替换与优化(含命名与索引重建)

10.1 生图模型建议(房间装修)

当用户使用"搬新家 / 找中介"时,优先推荐:

  1. gemini nanobanana pro
  2. gemini nanobanana 2

其他模型在房间结构保持与风格一致性上可能不稳定。

推荐配置:

  • GEMINI_API_KEY
  • GEMINI_MODEL=nanobanana-pro(或 nanobanana-2

并提示用户:缺 key 时可在侧边栏内直接填写并保存。

10.2 侧边栏验证码安全提醒(必须)

默认验证码为 1234,但生产/公网场景必须改强密码:

export ASSET_DRAWER_PASS="your-strong-pass"

理由:防止外部访问者修改房间布局、装饰和资产配置。

10.3 版权口径更新

主角状态素材已切换为无版权争议的小猫,不再沿用旧角色版权说明。

保留统一口径:

  • 代码:MIT
  • 美术资产:禁止商用

10.4 安装时必须提醒(API 可选)

在帮助主人安装时,需明确提醒:

  • 现在支持接入自己的生图 API 来改美术资产与背景(可持续更换)。
  • 但基础功能(状态看板、多 Agent、资产替换/布局、三语切换)不依赖 API,不开 API 也能正常使用。

建议对主人口径:

先把基础看板跑起来;需要"无限换背景/AI 生图装修"再接入自己的 API。

10.5 老用户更新指南(从旧版本升级)

如果主人之前已经下载过旧版,按以下步骤升级:

  1. 进入项目目录并备份本地配置(如 state.json、自定义资产)。
  2. 拉取最新代码(git pull 或重新克隆到新目录)。
  3. 确认依赖:python3 -m pip install -r backend/requirements.txt
  4. 保留并检查本地运行配置:
    • ASSET_DRAWER_PASS
    • GEMINI_API_KEY / GEMINI_MODEL(如需生图)
  5. 如有自定义位置,确认:
    • asset-positions.json
    • asset-defaults.json
  6. 重启后端并验收关键功能:
    • /health
    • 三语切换(CN/EN/JP)
    • 资产侧栏(选择、替换、设默认)
    • 生图入口(有 key 时可用)

10.6 功能更新提醒清单(对主人口播)

本次更新以后,至少提醒主人以下变化:

  1. 已支持 CN/EN/JP 三语切换(含 loading 与气泡实时联动)。
  2. 已支持 自定义美术资产替换(含动态素材切帧同步,减少闪烁)。
  3. 已支持 接入自有生图 API 持续更换背景(推荐 nanobanana-pro / nanobanana-2)。
  4. 新增/强化了安全项:ASSET_DRAWER_PASS 生产环境建议改强密码。

10.7 2026-03-05 稳定性修复

本次更新修复了多个影响线上稳定运行的问题:

  1. CDN 缓存修复:静态资源 404 不再被 CDN 长缓存(之前导致 phaser.js 被缓存为 404 达 2.7 天)。
  2. 前端加载修复:修复 fetchStatus() 中的 JS 语法错误(多余 else 块),解决页面卡 loading 问题。
  3. 生图异步化:生图接口改为后台任务 + 轮询模式,避免 Cloudflare 524 超时(100s 限制)。前端显示实时等待进度。
  4. 移动端侧边栏:新增遮罩层、body 滚动锁定、100dvh 适配、overscroll-behavior: contain
  5. Join Key 增强:支持 key 级别过期时间(expiresAt)和并发上限(maxConcurrent),join-keys.json 不再入库。

详细说明见:docs/UPDATE_REPORT_2026-03-05.md

Comments

Loading comments...