← 返回 Skills 市场
For long videos, automatically split the video task, using the last frame of the previous video as the first frame of the current video to maintain video continuity
作者
liusaikang
· GitHub ↗
· v1.0.1
· MIT-0
77
总下载
0
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install seedance-2-0-video-generation
功能描述
Generate AI videos using ByteDance Seedance 2.0. Use when the user wants to: (1) generate videos from text prompts, (2) generate videos from images (first fr...
安全使用建议
Before installing: (1) be aware this skill will upload any local images you pass (converted to base64) and send them to https://ark.cn-beijing.volces.com — do not upload sensitive images unless you trust that service and have the right to transmit them; (2) the skill requires an ARK_API_KEY (the SKILL.md and seedance.py enforce this), but the registry metadata does not declare it — confirm the manifest and supply the key only if you trust the operator; (3) there is no install spec even though SKILL.md references a CLI path — confirm where seedance.py will be stored and run, and inspect seedance.py yourself (it is included) to ensure there are no unexpected behaviors; (4) if you need higher assurance, ask the publisher for source/homepage, a clear install instruction, and explicit metadata listing ARK_API_KEY as required. If those inconsistencies are resolved (metadata updated and install path clarified), the skill would be coherent for its stated purpose.
功能分析
Type: OpenClaw Skill
Name: seedance-2-0-video-generation
Version: 1.0.1
The skill bundle provides a CLI tool (`seedance.py`) and instructions (`SKILL.md`) for generating videos via the Volcengine Ark API. A shell injection vulnerability exists in `seedance.py` within the `cmd_wait_logic` function, where a `task_id` returned from the remote API is used to construct a file path that is passed directly to `os.system` on macOS. If the API were compromised or malicious, it could return a crafted task ID to execute arbitrary commands. While the code lacks evidence of intentional malice and includes helpful safety instructions for the agent (e.g., requiring user confirmation for long-video generation), the lack of input sanitization on API-provided data constitutes a high-risk vulnerability.
能力评估
Purpose & Capability
The description, SKILL.md, and seedance.py consistently target ByteDance Seedance via the Volcengine Ark API (model doubao-seedance-2-0-260128). Requiring an API key for that service is expected, but the registry metadata lists no required environment variables or primary credential — this mismatch is incoherent and should be corrected.
Instruction Scope
SKILL.md and seedance.py stay within the stated purpose: creating/querying/deleting video generation tasks and optionally downloading results. However, the tool will read local image files (convert to base64) and upload them to the external API, and it will download generated videos/last frames to local directories — users should be aware that local files and generated content are transmitted to/from the external service.
Install Mechanism
There is no install spec (no instructions that place the CLI at the claimed path), yet SKILL.md refers to a CLI at ~/.claude/skills/seedance-2-0-video/seedance.py and the bundle includes seedance.py. Absence of an explicit install mechanism is not necessarily malicious but is an inconsistency to clarify (where will the script be placed/run?).
Credentials
The code and SKILL.md require ARK_API_KEY (Bearer token) to call Volcengine APIs — that credential is proportionate to the skill's purpose. The problem is the registry metadata does not declare this required environment variable or primary credential, which is misleading and could cause unexpected behavior or credential surprises at runtime.
Persistence & Privilege
The skill does not request permanent/always-on presence, does not modify other skills, and has no install script that changes system-wide configuration. It performs normal network calls and file I/O for its stated purpose.
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 seedance-2-0-video-generation - 安装完成后，直接呼叫该 Skill 的名称或使用
/seedance-2-0-video-generation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
v1.0.1
- Documentation updates only; no functional changes to the skill.
- SKILL.md reformatted or touched, but version and implementation remain unchanged.
v1.0.0
Initial release of Seedance 2.0 Video Generation skill.
- Generate AI videos from text prompts, images (first frame or first + last frame), or combine for more control.
- Supports chaining segments for long videos (>15s) with last-frame continuity.
- Audio generation and draft mode (cheaper, faster previews) are available.
- Python CLI tool provided for robust task creation, polling, and downloads; raw API/curl guides included.
- Manage video generation tasks (create, check status, download result) via Volcengine Ark Seedance 2.0 model.
元数据
常见问题
For long videos, automatically split the video task, using the last frame of the previous video as the first frame of the current video to maintain video continuity 是什么？
Generate AI videos using ByteDance Seedance 2.0. Use when the user wants to: (1) generate videos from text prompts, (2) generate videos from images (first fr... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 77 次。
如何安装 For long videos, automatically split the video task, using the last frame of the previous video as the first frame of the current video to maintain video continuity？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install seedance-2-0-video-generation」即可一键安装，无需额外配置。
For long videos, automatically split the video task, using the last frame of the previous video as the first frame of the current video to maintain video continuity 是免费的吗？
是的，For long videos, automatically split the video task, using the last frame of the previous video as the first frame of the current video to maintain video continuity 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
For long videos, automatically split the video task, using the last frame of the previous video as the first frame of the current video to maintain video continuity 支持哪些平台？
For long videos, automatically split the video task, using the last frame of the previous video as the first frame of the current video to maintain video continuity 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 For long videos, automatically split the video task, using the last frame of the previous video as the first frame of the current video to maintain video continuity？
由 liusaikang（@liusaikang）开发并维护，当前版本 v1.0.1。
推荐 Skills