← Back to Skills Marketplace
qwencloud-video-generation
by
Cuixiaoyang123
· GitHub ↗
· v0.1.1
· MIT-0
123
Downloads
0
Stars
0
Active Installs
2
Versions
Install in OpenClaw
/install qwencloud-video-generation
Description
[QwenCloud] Generate videos using Wan models. Supports text-to-video, image-to-video, first+last frame, reference-based role-play, and video editing (VACE)....
Usage Guidance
This skill appears to be a legitimate QwenCloud (DashScope) video-generation tool, but the registry metadata omits the DASHSCOPE_API_KEY that the scripts and SKILL.md clearly require. Before installing or running: 1) Confirm you are comfortable providing a QwenCloud API key (DASHSCOPE_API_KEY) and that you will store it in a safe place (prefer .env or environment variable, not command history). 2) Inspect the bundled scripts (scripts/video.py and qwencloud_lib.py) to verify they only call the documented dashscope endpoints and do not exfiltrate data elsewhere. 3) Be aware the skill can read local files you point it at (for uploads) and can load .env files in the repo or current directory — don't point it at sensitive files. 4) If the skill proposes to modify agent config files, refuse or verify the exact changes before agreeing. 5) Ask the publisher to update registry metadata to declare DASHSCOPE_API_KEY (and any other required env vars) so required permissions are transparent. If you want, run the scripts in an isolated environment (VM / container) with a throwaway API key first.
Capability Analysis
Type: OpenClaw Skill
Name: qwencloud-video-generation
Version: 0.1.1
The skill bundle implements telemetry and aggressive update-check mechanisms that leverage the AI agent to modify the host environment. Specifically, 'scripts/qwencloud_lib.py' contains logic to inspect the file system path and call stack to identify the active agent and skill, reporting this metadata to the API provider via the 'X-DashScope-Source-Config' header. Furthermore, 'SKILL.md' and 'scripts/gossamer.py' instruct the agent to monitor stderr for specific signals and prompt the user to execute shell commands ('npx skills add') to install additional software. While these features are framed as maintenance and analytics, the use of the agent to drive software installation and the collection of environment-specific metadata are high-risk behaviors.
Capability Tags
Capability Assessment
Purpose & Capability
The skill is clearly a QwenCloud (DashScope) video-generation integration and the bundled scripts and SKILL.md show expected API calls to dashscope-intl.aliyuncs.com. However, the registry metadata declares no required environment variables or primary credential, while SKILL.md and the included Python modules explicitly expect a DASHSCOPE_API_KEY (and optionally QWEN_BASE_URL/QWEN_REGION). This mismatch (missing declared credential) is incoherent and should be corrected or explained.
Instruction Scope
Runtime instructions are detailed and mostly limited to submitting async video tasks, polling, downloading, and local media post-processing. They also instruct the agent to read bundled scripts and reference docs, load .env files, upload user-provided local files, and (optionally, with user consent) append entries to agent config files. Those behaviors are expected for a video-generation skill, but they grant the skill the ability to read arbitrary paths referenced by the user and to modify agent config files if permitted — the SKILL.md asks to 'Ask the user before modifying any file', which mitigates but does not eliminate the risk.
Install Mechanism
No external install/download is declared; the skill is instruction-plus-bundled-scripts only. No remote install URLs or package installs appear in the manifest, lowering install-time risk.
Credentials
Functionally the skill needs an API key (DASHSCOPE_API_KEY) and may use QWEN_BASE_URL/QWEN_REGION and local .env files, which are proportional to contacting the provider and supporting custom endpoints. The problem is that the registry metadata lists no required env vars or primary credential, so the skill's declared requirements understate the secrets it will need and reference at runtime. The scripts also read .env and arbitrary local file paths (for uploads), which is appropriate for local-file inputs but should be made explicit in metadata so users know what sensitive information may be accessed.
Persistence & Privilege
The skill does not request always:true and does not appear to demand permanent system-wide privileges. It offers optional guidance to append a block to agent config files for agent compatibility, but the SKILL.md explicitly instructs to ask the user before modifying files. Autonomous invocation is allowed (platform default) but not combined with other alarming privileges.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install qwencloud-video-generation - After installation, invoke the skill by name or use
/qwencloud-video-generation - Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.1.1
- Added support for wan2.7 models (t2v and i2v) with advanced features such as ratio control, auto-dubbing, unified i2v protocol, and media array input.
- Updated mode selection and model tables to reflect wan2.7 model usage and required parameters.
- Clarified execution: explicitly document how to select new models, and direct users to official model detail pages for specific info.
- Enhanced documentation on i2v mode: now covers unified protocol (first/last frames, video continuation, audio sync) for wan2.7.
- No code or file-level changes detected in this release.
v0.1.0
Initial release of qwencloud-video-generation skill, enabling advanced video creation and editing with Wan models.
- Supports text-to-video, image-to-video, first+last frame transitions, reference-based role-play, and video editing (VACE).
- Detailed documentation included for mode selection, model compatibility, API key setup, and usage precautions.
- Provides step-by-step guidance for script execution, polling, media merging, and troubleshooting.
- Emphasizes strict security: never outputs API keys or plaintext credentials; guides users to set environment variables correctly.
- Outlines differences in parameters and capabilities by mode, and includes a model/version selection guide.
- Clarifies triggers for using this skill vs. related QwenCloud skills (image generation, video understanding).
Metadata
Frequently Asked Questions
What is qwencloud-video-generation?
[QwenCloud] Generate videos using Wan models. Supports text-to-video, image-to-video, first+last frame, reference-based role-play, and video editing (VACE).... It is an AI Agent Skill for Claude Code / OpenClaw, with 123 downloads so far.
How do I install qwencloud-video-generation?
Run "/install qwencloud-video-generation" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is qwencloud-video-generation free?
Yes, qwencloud-video-generation is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does qwencloud-video-generation support?
qwencloud-video-generation is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created qwencloud-video-generation?
It is built and maintained by Cuixiaoyang123 (@cuixiaoyang123); the current version is v0.1.1.
More Skills