← 返回 Skills 市场
volcengine-skills

Byted Las Video Resize

作者 volcengine-skills · GitHub ↗ · v1.0.1 · MIT-0
cross-platform ⚠ suspicious
137
总下载
0
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install byted-las-video-resize
功能描述
Resizes, scales, and adjusts video resolution and dimensions using GPU-accelerated NVENC encoding via Volcengine LAS. Video resizing, video scaling, video up...
使用说明 (SKILL.md)

LAS 视频分辨率调整(las_video_resize

将视频调整到指定分辨率范围,支持宽高比保持策略和 GPU NVENC 编码。

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_video_resize
  • API: 异步(submitpoll
  • 环境变量: LAS_API_KEY (必填)

详细参数与接口定义见 references/api.md

Gotchas

  • 强制固定尺寸:需要设置 min_width=max_widthmin_height=max_height,并将 force_original_aspect_ratio_type=disable
  • 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: 提交任务
- [ ] Step 4: 异步查询
- [ ] Step 5: 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. 检查输出路径
    • output_tos_dir 为必填参数,必须由用户提供自己可写的 TOS 目录路径(格式:tos://bucket/output_dir/)。
    • 服务端需要将缩放后的视频写入此目录。
  4. 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 "$0")/scripts/env_init.sh" las_video_resize
workdir=$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:使用 File API 上传(只需 LAS_API_KEY,无需 TOS 凭证和 Bucket):
    lasutil file-upload \x3Clocal_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. 获取视频信息(时长、编码、分辨率):
    lasutil media-duration \x3Cvideo_url>
    
  3. 根据档位查表计算总价,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: 提交任务 (Submit)

构造基础 data.json

{
  "video_path": "\x3Cpresigned_url>",
  "output_tos_dir": "tos://\x3Cyour-bucket>/output_dir/",
  "output_file_name": "resized.mp4",
  "min_width": 1280, "max_width": 2560,
  "min_height": 720, "max_height": 1440,
  "force_original_aspect_ratio_type": "increase"
}

重要提示: output_tos_dir 必须由用户提供,需要填写用户自己账号下可写的 TOS 目录(服务端会将缩放后的视频写入此目录)。

单文件提交

data=$(cat "$workdir/data.json")
lasutil submit las_video_resize "$data" > "$workdir/submit.json"

⚠️ 强制反馈:任务提交成功后,必须立即向用户返回生成的 task_id,以便用户跟踪进度或在必要时手动查询。

Step 4: 异步查询 (Poll)

⚠️ 异步任务与后台轮询约束

  • 如果你当前的环境支持后台任务/异步长效运行:你可以利用环境提供的后台能力(例如发起后台轮询任务),并在任务完成后主动将结果返回给用户。
  • 如果你当前的环境不支持长效后台任务(如普通的单轮对话沙箱),且直接 sleep 循环会导致超时崩溃:绝对禁止在代码中执行死循环等待! 此时必须立即向用户输出 Task ID 并结束当前轮次,告知用户:"任务已提交,请稍后向我询问进度"。

单任务查询

lasutil poll las_video_resize {task_id}
  • COMPLETED → 返回输出视频路径 result.data.output_path
  • RUNNING/PENDING → 稍后重试。

Step 5: 结果呈现

处理结果

# 获取缩放后的视频 URL
resized_url=$(cat "./output/{task_id}/result.json" | jq -r '.data.output_path')
echo "缩放后的视频: $resized_url"

视频文件

  • 缩放后的视频已保存在 TOS,直接返回预签名 URL
  • 无需再次上传

向用户展示

  1. 缩放后的视频下载链接
  2. 输出分辨率、编码信息
  3. 本地结果路径:./output/{task_id}/
  4. 计费声明

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
安全使用建议
This skill appears to implement Volcengine LAS video resizing but has several red flags you should address before running it: - Verify the missing metadata: SKILL.md requires LAS_API_KEY / LAS_REGION but the registry lists no required env vars — expect to supply LAS_API_KEY. Ask the publisher to clarify and add a homepage or official source. - Inspect env_init.sh before sourcing: it downloads a manifest and pip-installs a wheel from a remote URL. Do not run this script unless you trust that host. Prefer installing the official SDK from a verified package index or vendor-provided URL. - Avoid allowing the scripts to source your entire .env. If you must provide LAS_API_KEY, set it in a minimal, dedicated location (or export it just for the command) instead of revealing other secrets. - The repo contains ASR-related leftovers (transcript handling), which suggests careless copying. Test the skill in an isolated sandbox or VM before using it with production credentials or sensitive data. - If you need this functionality but are unsure of the origin, ask the publisher for provenance (homepage, repo, or organization), a signed release, or a verified pip package name; otherwise consider implementing a minimal wrapper that only calls the official LAS API endpoints you trust. If you decide to proceed, run scripts only after manual review, and keep LAS_API_KEY scoped to the minimal permissions required (and rotate it afterward if you suspect exposure).
功能分析
Type: OpenClaw Skill Name: byted-las-video-resize Version: 1.0.1 The skill bundle is designed for video resizing using Volcengine LAS and appears to be legitimate. It includes scripts for environment setup (`env_init.sh`), format validation, and background task polling. While `env_init.sh` downloads and installs a Python wheel from a remote URL, the domain (volces.com) is the official Volcengine cloud platform. A minor bug was identified in `scripts/generate_result.md.sh`, which contains placeholder text and logic for ASR (speech-to-text) tasks instead of video resizing, likely due to a copy-paste error from a different skill template. No evidence of malicious intent, data exfiltration, or harmful prompt injection was found.
能力标签
requires-sensitive-credentials
能力评估
Purpose & Capability
The SKILL.md clearly requires LAS_API_KEY (and optionally LAS_REGION / VOLCENGINE_ACCESS_KEY & SECRET for certain flows), but the registry metadata lists no required env vars — a mismatch. Requiring TOS output paths and TOS credentials is reasonable for a service that writes back to user buckets, but the skill also contains files (generate_result.md.sh) that reference ASR/transcript outputs, which does not match a pure video-resize purpose and suggests copy/paste or leftover artifacts.
Instruction Scope
The runtime instructions tell the agent to source scripts/env_init.sh (which fetches a remote manifest and may pip-install a wheel), to read/supply LAS_API_KEY (including asking user to create env files), to upload local files via lasutil, submit tasks, and optionally run a background poll script. Sourcing a project .env is explicitly performed by scripts; that can expose unrelated secrets if a user .env exists. The requirement that the user provide an output_tos_dir and the explicit step-by-step workflow are coherent, but the instruction set gives the agent permission to run network downloads and execute installs — a significant expansion of scope beyond mere API calls.
Install Mechanism
There is no declared install spec, but scripts/env_init.sh fetches a remote manifest via curl and pip-installs a wheel from https://las-ai-cn-beijing-online.tos-cn-beijing.volces.com/operator_cards_serving/public/skills/sdk/las_sdk-0.2.0-py3-none-any.whl. That is a dynamic remote download and install (archive execution) performed at runtime when the agent sources the script. Even if the host appears related to Volcengine, installing a wheel from an external URL at runtime is a higher-risk pattern and should be verified.
Credentials
SKILL.md requires LAS_API_KEY and LAS_REGION (and notes that VOLCENGINE_ACCESS_KEY / VOLCENGINE_SECRET may be needed for certain output-download flows), but the registry metadata claims no required env vars — inconsistent. The scripts will source a .env file if present, which can unintentionally load unrelated secrets into the agent environment. The number of env variables requested is small and appropriate for the service, but the practice of sourcing an entire project .env and asking users to place keys in env.sh is risky and should be limited to the minimum-required variables only.
Persistence & Privilege
always:false (normal). The skill's scripts create/activate a local virtual environment (.las_venv) and create a temporary LAS_WORKDIR, and may pip-install packages into the venv; this writes to disk and installs code, but is scoped to the skill's directory/venv. The skill does not request always:true and does not modify other skills' config. The main concern is the runtime ability to install and run arbitrary code fetched from the network.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install byted-las-video-resize
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /byted-las-video-resize 触发
  4. 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
v1.0.1
byted-las-video-resize v1.0.1 - Changed skill name and description from audio conversion to GPU-accelerated video resizing/scaling with async batch workflow and detailed user guidance. - Added scripts for environment initialization, file format checking, result generation, and async polling to streamline process automation. - Introduced documentation on pricing and evaluation flows; added checklists for stepwise workflow compliance. - Updated references and API parameters to match video resizing use-cases. - Removed legacy audio conversion script.
v1.0.0
- Initial release of the audio format conversion skill. - Supports converting audio files between wav, mp3, and flac formats. - Allows changing audio properties such as sample rate and bit rate with ffmpeg parameters. - Handles input and output files stored in TOS. - Requires setting the LAS_API_KEY for authentication.
元数据
Slug byted-las-video-resize
版本 1.0.1
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 2
常见问题

Byted Las Video Resize 是什么?

Resizes, scales, and adjusts video resolution and dimensions using GPU-accelerated NVENC encoding via Volcengine LAS. Video resizing, video scaling, video up... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 137 次。

如何安装 Byted Las Video Resize?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install byted-las-video-resize」即可一键安装,无需额外配置。

Byted Las Video Resize 是免费的吗?

是的,Byted Las Video Resize 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Byted Las Video Resize 支持哪些平台?

Byted Las Video Resize 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 Byted Las Video Resize?

由 volcengine-skills(@volcengine-skills)开发并维护,当前版本 v1.0.1。

💬 留言讨论