Skill flagged — suspicious patterns detected

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

Byted Las Image Resample

v1.0.1

Downsamples, resizes, and compresses images using Volcengine LAS with multiple interpolation methods (lanczos, bilinear, nearest). Image resizing, image scal...

0· 117·0 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 volcengine-skills/byted-las-image-resample.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "Byted Las Image Resample" (volcengine-skills/byted-las-image-resample) from ClawHub.
Skill page: https://clawhub.ai/volcengine-skills/byted-las-image-resample
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 byted-las-image-resample

ClawHub CLI

Package manager switcher

npx clawhub@latest install byted-las-image-resample
Security Scan
Capability signals
Requires sensitive credentials
These labels describe what authority the skill may exercise. They are separate from suspicious or malicious moderation verdicts.
VirusTotalVirusTotal
Suspicious
View report →
OpenClawOpenClaw
Suspicious
medium confidence
!
Purpose & Capability
The SKILL.md and scripts clearly require a Volcengine API key (LAS_API_KEY), may require LAS_REGION, and in some flows request VOLCENGINE_ACCESS_KEY/VOLCENGINE_SECRET_KEY for TOS downloads. They also assume availability of lasutil and ImageMagick's identify. However the registry metadata lists no required environment variables or binaries. This mismatch is incoherent: an image-resample skill legitimately needs a LAS API key and lasutil, but the metadata should declare them; the absence suggests sloppy or incomplete packaging.
!
Instruction Scope
Runtime instructions tell the agent to 'source' scripts/env_init.sh, run lasutil file-upload/process, and use identify to inspect local images. The env_init script fetches a remote manifest and may auto-upgrade the SDK. Instructions also tell the agent to prompt the user for missing keys and to upload local files to TOS. The steps are mostly consistent with the stated image-resample purpose, but the auto-update/install behavior and the assumption that identify/lasutil exist are out-of-band relative to the registry metadata and widen the skill's runtime scope.
!
Install Mechanism
There is no formal install spec, but scripts/env_init.sh performs network actions: curl to a manifest URL and pip install of a wheel from https://las-ai-cn-beijing-online.tos-cn-beijing.volces.com/.../las_sdk-0.2.0-py3-none-any.whl. Installing a remote wheel without a checksum or well-known release host is higher risk (arbitrary code can be executed). The domain appears to be a cloud storage endpoint (TOS/volces), not a mainstream package registry; automatic upgrade behavior runs without explicit user approval in the script flow.
!
Credentials
The skill reasonably needs LAS_API_KEY and may require LAS_REGION and, in some flows, VOLCENGINE_ACCESS_KEY/VOLCENGINE_SECRET_KEY when placing outputs on TOS. Those are sensitive credentials and should be declared by the skill metadata; they are not. The SKILL.md also suggests creating an env.sh file with exports for keys, which increases the chance users will place secrets in repo-adjacent files. Requesting these credentials is proportionate to the functionality, but the missing metadata declaration and the instruction to create local env files are concerning.
Persistence & Privilege
The skill does not request always:true and does not modify other skills' configs. However env_init.sh creates/uses a virtual environment under the project (writing files to disk) and pip-installs a remote wheel into that environment. That gives the skill persistent code on disk and the ability to run arbitrary Python code (via the installed SDK). This is expected for a tool wrapper but increases blast radius and should be done with explicit user consent and verification of the remote package.
What to consider before installing
This skill mostly does what it claims (it wraps Volcengine LAS via lasutil), but there are important red flags you should address before installing or running it: - Metadata mismatch: The registry entry lists no required env vars or binaries, but SKILL.md and scripts require LAS_API_KEY, LAS_REGION, and tools like lasutil and ImageMagick (identify). Treat that as a packaging bug and verify requirements manually. - Remote install: The startup script will curl a manifest and pip-install a wheel from a cloud URL (not a standard public release host) without checksum. Do not run this in a sensitive environment. Inspect the wheel URL, prefer an official SDK package or request a signed checksum, and run in an isolated environment if you proceed. - Secrets handling: The skill asks for LAS credentials and may ask for VOLCENGINE_ACCESS_KEY/SECRET_KEY when downloading results. Do not paste long-lived keys into chat. Prefer scoped, least-privilege credentials and avoid storing secrets in plain env.sh files in shared directories. - Local tools: Ensure lasutil and ImageMagick are available and from trusted sources. If they are not present, the skill will fail or attempt remote fallbacks. If you want to use this skill safely: (1) ask the author to update registry metadata to declare required env vars and binaries; (2) request a reproducible, signed SDK distribution (or the wheel's checksum) and prefer installing it offline or from a trusted registry; (3) run env_init.sh only in an isolated test environment and review its network targets; (4) only provide the minimum credentials necessary and rotate them after use.

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

latestvk978q0jn9s9q4bex63d1ew1mbs850aj7
117downloads
0stars
2versions
Updated 1w ago
v1.0.1
MIT-0

LAS 图像重采样(las_image_resample

对图像进行降采样缩放(仅缩小,不支持放大),支持修改 DPI、JPG/PNG 互转。

设计模式

本 skill 主要采用:

  • Tool Wrapper:封装 lasutil CLI 调用
  • Pipeline:包含 Step 0 → Step N 的顺序工作流

核心 API 与配置

  • 算子 ID: las_image_resample
  • API: 同步(process
  • 环境变量: LAS_API_KEY (必填)

详细参数与接口定义见 references/api.md

Gotchas

  • 仅支持降采样target_size 不得超过原图尺寸,本算子不支持放大。
  • 输入限制:仅支持 JPEG/PNG/TIFF,≤100MB,像素总数 ≤225,000,000。
  • 密钥安全:若聊天框屏蔽密钥,让用户在当前目录创建 env.sh 并写入 export LAS_API_KEY="...",SDK 会自动读取。
  • 免责声明:最终回复结果时必须包含:"本方式的计费均为预估计费,与实际费用有差距,实际费用以运行后火山产生的账单为准。计费说明请参考 Volcengine LAS 定价。",且禁止使用"实际费用"字眼描述预估价。

工作流(严格按步骤执行)

复制此清单并跟踪进度:

执行进度:
- [ ] Step 0: 前置检查
- [ ] Step 1: 初始化与准备
- [ ] Step 2: 预估价格
- [ ] Step 3: 执行/提交任务
- [ ] Step 4: 结果呈现

Step 0: 前置检查(⚠️ 必须在第一轮对话中完成)

在接受用户的任务后,不要立即开始执行,必须首先进行以下环境检查:

  1. 检查 LAS_API_KEYLAS_REGION:确认环境变量或 .env 中是否已配置。
    • 若无,必须立即向用户索要(提示:LAS_REGION 常见为 cn-beijing)。
    • 注意LAS_REGION 必须与您的 API Key 及 TOS Bucket 所在的地域完全一致。如果用户中途切换了 Region,必须提醒用户其 TOS Bucket 也需对应更换,否则会导致权限异常或上传失败。
  2. 检查输入路径
    • 如果用户要求处理的是本地文件,则需要先通过 File API 上传至 TOS(只需 LAS_API_KEY,无需额外 TOS 凭证)。
    • 如果算子的输出结果存放在 TOS 上,且用户需要下载回本地,则需要 VOLCENGINE_ACCESS_KEYVOLCENGINE_SECRET_KEY。对于仅需要上传输入文件的场景,TOS 凭证不再必须
  3. 确认无误后:才能进入下一步。

Step 1: 初始化与准备

环境初始化(Agent 必做)

# 执行统一的环境初始化与更新脚本(会自动创建/激活虚拟环境,并检查更新)
source "$(dirname "$0")/scripts/env_init.sh" las_image_resample
workdir=$LAS_WORKDIR

如果网络问题导致更新失败,脚本会跳过检查,使用本地已安装的 SDK 继续执行。

  • 处理本地文件时:先本地检查格式和尺寸,预估价格,用户确认后再上传:
    # 提前检查图片格式(避免参数错误)
    ./scripts/check_format.sh <local_path>
    # 本地使用 identify 获取尺寸(无需上传即可预估价格)
    identify -format "%wx%h" <local_path>
    
    计算预估价格并等待用户确认后,再执行上传:
    # 用户确认后,上传到 TOS
    lasutil file-upload <local_path>
    
    上传成功后返回 JSON,取其中的 presigned_url(HTTPS 预签名下载链接,24 小时有效)传给算子作为输入 URL。

Step 2: 预估价格(⚠️ 必须获得用户确认)

  1. 读取 references/prices.md 获取最新计费标准。
  2. 优先本地获取尺寸(避免不必要上传):
    # 使用 ImageMagick identify 本地获取
    identify -format "%wx%h" <local_path>
    
    如果 identify 失败,再使用 lasutil 远程获取:
    lasutil image-info <input_image_url>
    
  3. 根据分辨率档位和单价计算总价,将计费单价与预估总价一并告知用户并强制暂停执行,明确等待用户回复确认。在用户明确回复"继续"、"确认"等同意指令前,绝对禁止进入下一步(执行/提交任务)。提示:预估仅供参考,实际以火山账单为准。计费说明请参考 Volcengine LAS 定价

Step 3: 执行重采样 (Process)

构造基础 data.json

{
  "image_src_type": "image_url",
  "image": "<presigned_url>",
  "tos_dir": "tos://bucket/output/",
  "target_size": [1024, 1024],
  "method": "lanczos"
}

注:本地文件通过 lasutil file-upload 上传后,使用 presigned_url 作为 image 字段值,image_src_type 设为 "image_url"。若直接使用 TOS 路径,设置 "image_src_type": "image_tos"

执行命令:

data=$(cat "$workdir/data.json")
lasutil process las_image_resample "$data" > "$workdir/result.json"

结果呈现

使用脚本自动生成结果展示(自动包含计费声明):

./scripts/generate_result.md.sh $workdir/result.json <estimated_price>

生成内容包含:

  • 任务信息卡片
  • 输入/输出尺寸统计
  • 自动包含计费声明

手动提取方式:

jq -r '"输出路径:" + .data.image_path' $workdir/result.json

审查标准

执行完成后,Agent 应自检:

  1. 环境变量是否正确配置
  2. 输入文件是否成功上传
  3. 输出结果是否正确呈现给用户
  4. 计费声明是否包含

Comments

Loading comments...