Skill flagged — suspicious patterns detected

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

腾讯云 COS 原生操作技能

v1.1.0

面向 OpenClaw 的腾讯云 COS 原生操作技能。适用于上传本地文件、批量同步目录、大文件分片上传、生成临时签名链接、浏览对象与文件夹视图、复制/移动/重命名/删除对象、统计目录体量,以及排查 COS 凭证和 Bucket 配置。

0· 115·0 current·0 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 156554395/tx-cos.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "腾讯云 COS 原生操作技能" (156554395/tx-cos) from ClawHub.
Skill page: https://clawhub.ai/156554395/tx-cos
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Required binaries: node, npm
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 tx-cos

ClawHub CLI

Package manager switcher

npx clawhub@latest install tx-cos
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
medium confidence
Purpose & Capability
The scripts implement Tencent COS operations (upload, list, delete, multipart upload, signed URLs, diagnostics) that align with the skill description. The required runtime capabilities (reading local files, calling COS SDK) are appropriate for the stated purpose.
Instruction Scope
SKILL.md and the scripts limit actions to COS API calls and local filesystem reads for files explicitly supplied to the scripts. The scripts only access environment variables prefixed COS_ and do not call external endpoints other than COS. They will, however, read any local file paths passed in, so the agent/run-time must not be instructed to upload sensitive local files unintentionally.
!
Install Mechanism
There is no install spec (no npm install step or dependency list). The scripts require the npm package 'cos-nodejs-sdk-v5' but the skill does not declare how/when that dependency is installed. That mismatch can cause runtime failures and is an operational/integrity concern (missing dependency declaration).
!
Credentials
The code requires COS_SECRET_ID, COS_SECRET_KEY, COS_REGION, COS_BUCKET (and optionally COS_DOMAIN) to function, which is proportionate to the purpose. However the registry metadata lists no required env vars — an inconsistency that could lead to the agent exposing or mismanaging credentials. The skill therefore implicitly needs sensitive secrets but does not declare them in metadata.
Persistence & Privilege
The skill does not request persistent or system-wide privileges (always is false). It does not modify other skills or global configuration. Autonomous invocation is allowed by default but not combined with other high-risk flags.
What to consider before installing
This skill appears to do what it says (manage Tencent COS objects) and the scripts are readable, but two mismatches raise caution: (1) the code requires COS_SECRET_ID, COS_SECRET_KEY, COS_REGION and COS_BUCKET at runtime but the registry metadata declares no required env vars — ask the publisher to add these to requires.env so you know when the skill needs secrets; (2) the scripts depend on 'cos-nodejs-sdk-v5' but there is no install spec — ensure the runtime will install or provide that package (or run npm install in the skill directory) before use. Only provide COS credentials to this skill if you trust the author and the runtime environment; limit the agent's ability to access unrelated local files (don't let it upload arbitrary system files). If anything looks unexpected, request the publisher add an explicit install section and manifest the needed env vars and a checksumed dependency list.

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

Runtime requirements

Binsnode, npm
latestvk9733na5pf0r9tfea8gxbq4s3d844v7j
115downloads
0stars
5versions
Updated 3w ago
v1.1.0
MIT-0

腾讯云 COS 技能

把腾讯云 COS 变成 OpenClaw 可直接调用的一组高频对象存储工作流。

适合这些典型任务:

  • 把本地文件或整个目录快速上传到 COS
  • 为图片、压缩包、构建产物生成临时访问链接
  • 浏览某个前缀下的文件结构,做复制、移动、重命名、清理
  • 处理视频、模型、备份包等大文件上传
  • 快速判断凭证、地域、Bucket、自定义域名哪里配错了

默认做法是:

  • 优先检查关键配置与目标路径
  • 在需要时生成并执行简洁脚本,直接完成对象存储操作
  • 结果以结构化 JSON 或简洁中文返回给用户

适用场景

当用户出现这些意图时启用:

  • 把本地文件上传到腾讯云 COS
  • 批量上传一个目录或多个文件到 COS
  • 上传大文件并控制并发、分片大小、断点续传策略
  • 获取 COS 对象的临时签名 URL
  • 列出某个前缀下的文件或文件夹
  • 创建、删除、移动、复制、重命名 COS 对象
  • 统计某个目录下的文件数量和总大小
  • 排查 COS 凭证、地域、Bucket、自定义域名配置

前置条件

要求具备可访问腾讯云 COS 的运行环境和必要配置。

如果关键配置不完整,先明确缺什么,不要猜。

脚本资源

优先使用 scripts/ 里的可执行脚本,而不是每次临时重写一遍逻辑。

可直接复用的脚本:

  • scripts/check_config.js - 检查关键配置是否齐全
  • scripts/upload_file.js - 上传单个文件
  • scripts/upload_batch.js - 批量上传文件或整个目录
  • scripts/generate_signed_url.js - 生成临时签名链接
  • scripts/list_objects.js - 列出对象或文件夹视图
  • scripts/manage_objects.js - 删除、批量删除、复制、移动、重命名、创建文件夹
  • scripts/upload_large_file.js - 大文件分片上传
  • scripts/delete_folder.js - 删除文件夹,支持递归删除
  • scripts/get_folder_stats.js - 统计目录文件数量、总体积和类型分布

调用方式统一为:

node projects/skills/tx-cos/scripts/<script>.js '<json>'

默认执行策略

1. 先检查配置

优先先跑 scripts/check_config.js 确认关键配置是否齐全。

如果必需项缺失,直接告诉用户缺哪几个,不继续执行上传/删除类操作。

2. 需要真实操作时,优先调用现成脚本

这个技能的目标是让代理直接完成对象存储相关工作,而不是额外引入中间层。

优先采用脚本化、输入输出明确的执行方式;只有脚本覆盖不了时,才补临时逻辑。

3. 输出要面向人类

返回结果时优先包含:

  • 操作是否成功
  • 对象键 / 路径
  • URL / ETag / 大小 / 数量等关键信息
  • 失败时给出可执行的修复建议

不要把整段 SDK 原始响应直接甩给用户,除非对方明确要原始 JSON。

常用工作流

上传单文件

适用:用户说“把这个文件传到 COS”。

执行前要先确认本地文件存在、路径是文件而不是目录。返回结果时至少包含对象键、访问地址或资源标识、以及是否上传成功。

批量上传

适用:上传整个目录、多个文件、构建产物。

建议策略:

  • 先枚举文件列表
  • 明确远端前缀映射规则
  • 控制并发,避免一下子打爆带宽
  • 汇总成功数、失败数、失败文件

如果用户没有指定对象键规则,默认保留相对路径。

大文件上传

适用:视频、模型文件、压缩包。

优先用分片上传接口,并显式设置:

  • SliceSize
  • AsyncLimit
  • 进度回调

如果要做长时间上传,告诉用户:

  • 文件大小
  • 并发设置
  • 预计风险(网络抖动、覆盖同名对象)

获取签名 URL

适用:用户要临时下载链接、分享给别人。

关键参数:

  • Key
  • Expires

默认有效期可设为 3600 秒;如果用户没说,明确说明是临时链接。

列目录 / 文件夹视图

COS 没有真实目录,通常要用 Prefix + Delimiter: '/' 模拟文件夹。

当用户要“看看某个目录里有什么”时:

  • 优先返回文件夹和文件的分组视图
  • 结果过多时只展示前若干项,并提示可继续翻

删除、复制、移动、重命名

这些操作要更谨慎:

  • 删除前,确认对象键或前缀范围
  • 批量删除前,最好先列出将被删除的对象数量
  • 移动/重命名本质通常是复制后删除,要提示这一点
  • 跨 bucket 复制时,确认目标 bucket 是否由当前凭证可访问

如果用户表达不明确,比如“删一下旧文件”,先澄清范围。

安全边界

  • 绝不在回复里泄露敏感凭证
  • 输出配置检查结果时,只给出必要状态,不展示敏感值
  • 对删除、批量删除、递归删除操作保持高警惕
  • 如果涉及公开分享链接,提醒用户签名 URL 有时效,公开 Bucket 需额外确认
  • 不要擅自假设 Bucket 名、地域、对象路径

故障排查

常见问题与处理:

  • 凭证错误:检查访问凭证是否正确
  • 地域错误:检查地域配置是否和 Bucket 实际地域一致
  • Bucket 名错误:确认名称是否完整且可访问
  • 本地路径不存在:先校验文件存在
  • 自定义域名访问异常:先回退到默认访问域名验证
  • 大文件失败:降低并发、增大分片、重试上传

Comments

Loading comments...