← Back to Skills Marketplace
107
Downloads
0
Stars
0
Active Installs
2
Versions
Install in OpenClaw
/install openclaw-video-pipeline
Description
OpenClaw two-step pipeline: (1) expand a user brief into a three-shot storyboard and save it to storyboard/storyboard.json at project root, (2) run python3 s...
Usage Guidance
This skill appears to do what it says: produce a storyboard JSON and call the Zhipu (open.bigmodel.cn) video-generation API three times. Before installing/using it: (1) confirm and supply ZHIPUAI_API_KEY in your environment (the metadata omitted this requirement), (2) do not commit your API key to the repo (SKILL.md already warns this), (3) be aware the agent will write storyboard/storyboard.json at the project root and may download three MP4s to an output directory you specify, and (4) verify you trust the remote API (open.bigmodel.cn) since prompts and generated media are sent/received over the network. If you need higher assurance, request the maintainer update the package metadata to declare the required env var and perform a code review in a safe/test repository first.
Capability Analysis
Type: OpenClaw Skill
Name: openclaw-video-pipeline
Version: 1.0.1
The skill provides a legitimate two-step pipeline for generating 30-second videos using the Zhipu CogVideoX-3 API. The Python script (scripts/video-generate.py) uses standard libraries to handle asynchronous API requests, polling, and optional file downloads, while the SKILL.md provides clear instructions for the agent to manage the storyboard JSON. No evidence of malicious intent, data exfiltration, or risky command execution was found.
Capability Assessment
Purpose & Capability
The skill name/description (three 10s segments via CogVideoX-3) aligns with the included script and SKILL.md: the script reads a storyboard JSON and calls the open.bigmodel.cn async video API three times. However, the registry metadata lists no required environment variables or primary credential while the runtime instructions and script require ZHIPUAI_API_KEY — this metadata omission is an inconsistency.
Instruction Scope
SKILL.md clearly limits behavior to (1) writing storyboard/storyboard.json in the project root and (2) running scripts/video-generate.py to call the CogVideoX-3 REST API. It does not instruct reading unrelated system files or sending data to unexpected endpoints; network calls go to open.bigmodel.cn as documented. Note: the agent will create files in the repository root (storyboard/ and optional downloaded MP4s).
Install Mechanism
There is no install spec (instruction-only with a single Python script). The script uses only Python stdlib (with optional certifi) and performs HTTPS requests; no remote code downloads or archive extraction are present.
Credentials
At runtime the script requires ZHIPUAI_API_KEY (and optionally BIGMODEL_API_BASE). That credential is proportional to the described functionality. The concern is that the skill registry metadata did not declare this required environment variable or primary credential, which is a mismatch that could lead to unexpected failures or misconfiguration and should be corrected/confirmed before use.
Persistence & Privilege
The skill does not request permanent installation, does not set always:true, and does not modify other skills or system-wide settings. It runs on-demand and has typical network access for an API-backed task.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install openclaw-video-pipeline - After installation, invoke the skill by name or use
/openclaw-video-pipeline - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
**Added default storyboard file path and directory for improved workflow:**
- Now saves the storyboard JSON to storyboard/storyboard.json in the project root.
- video-generate.py will, by default, read from storyboard/storyboard.json if no input is specified.
- Added directory creation guidance for storyboard/.
- Documentation updated to clarify new default paths and simplify recommended usage.
- No changes to API or script logic; only file outputs and documentation updated for consistency and ease of use.
v1.0.0
OpenClaw video pipeline initial release:
- Introduces a two-step pipeline for 30-second videos: expands a user brief into a 3-shot storyboard, then generates 3 video segments using Zhipu CogVideoX-3.
- Provides detailed JSON schema and example prompts to ensure visual and narrative consistency across three 10-second video segments.
- Includes instructions for running generation scripts and setting required environment variables.
- Requires manual post-processing to concatenate three resulting videos into a single 30s clip.
- Links to official CogVideoX-3 documentation for further guidance.
Metadata
Frequently Asked Questions
What is video 生成流水线?
OpenClaw two-step pipeline: (1) expand a user brief into a three-shot storyboard and save it to storyboard/storyboard.json at project root, (2) run python3 s... It is an AI Agent Skill for Claude Code / OpenClaw, with 107 downloads so far.
How do I install video 生成流水线?
Run "/install openclaw-video-pipeline" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is video 生成流水线 free?
Yes, video 生成流水线 is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does video 生成流水线 support?
video 生成流水线 is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created video 生成流水线?
It is built and maintained by King (@ilaus); the current version is v1.0.1.
More Skills