← Back to Skills Marketplace
(Google) Veo 3 Video Gen
by
mango_fish
· GitHub ↗
· v1.0.7
· MIT-0
342
Downloads
0
Stars
1
Active Installs
8
Versions
Install in OpenClaw
/install veo-video-gen
Description
Generate videos using Pixwith API's Veo 3.1 model. Supports text-to-video and image-to-video (start/end frames). Two tiers: Fast (quick preview) and Pro (HD...
Usage Guidance
This skill appears to implement the Pixwith Veo video-generation API and asks for a PIXWITH_API_KEY, which is reasonable. However, the published registry metadata omitted that required API key — ask the publisher or registry maintainer to correct the metadata before installing. Confirm you trust https://pixwith.ai and are willing to send any images you provide to Pixwith/S3 (presigned uploads expire in ~10 minutes). Store the PIXWITH_API_KEY in your platform's secret manager (not in plain chat) and verify network permissions only allow the listed domains if your platform supports domain-restricted network access. As a precaution, test with non-sensitive, low-cost (Fast tier) jobs first and verify responses look legitimate. If the publisher cannot explain the metadata mismatch, treat the package as untrusted and do not provide sensitive images or credentials.
Capability Analysis
Type: OpenClaw Skill
Name: veo-video-gen
Version: 1.0.7
The skill provides a legitimate interface for AI video generation using the Pixwith Veo 3.1 API. It defines standard workflows for text-to-video and image-to-video tasks, utilizing curl for API requests and image uploads to S3-compatible storage. Permissions are appropriately restricted to the necessary API domains (api.pixwith.ai, uploads.pixwith.ai, and *.amazonaws.com) and user-provided image files, with no evidence of malicious intent, data exfiltration, or unauthorized system access.
Capability Assessment
Purpose & Capability
The SKILL.md describes a text-to-video and image-to-video integration with Pixwith Veo 3.1 and the network hosts it needs (api.pixwith.ai, uploads.pixwith.ai, and S3) — that matches the stated purpose. However, the registry metadata provided with the skill claims 'Required env vars: none' and 'Primary credential: none' while the SKILL.md explicitly requires PIXWITH_API_KEY. This mismatch is unexpected and reduces trust.
Instruction Scope
Runtime instructions are detailed and stay within video-generation scope: checking credits, creating tasks, polling status, obtaining presigned upload data, and uploading images via presigned S3 form fields. The skill asks the agent to read user-supplied local image files (explicitly allowed) and to post them to presigned upload URLs. It also tells the agent to treat API response tokens as opaque and reuse them exactly. Nothing in the instructions asks the agent to access unrelated local files, other credentials, or to exfiltrate arbitrary data, but the agent will transmit any uploaded user images to third-party storage (Pixwith/S3).
Install Mechanism
There is no install spec and no code files; this is instruction-only, so nothing will be written to disk by an installer. That minimizes install-time risk.
Credentials
The SKILL.md requires a single service credential (PIXWITH_API_KEY), which is proportional for a third-party API integration. The problem is the registry metadata omitted this requirement. That inconsistency is important: if the platform does not surface or protect PIXWITH_API_KEY appropriately, the skill will fail or the runtime might prompt for credentials unexpectedly. The skill also requests network access to *.amazonaws.com for presigned uploads — expected for S3-backed uploads but worth noting since it permits access to many hosts under that domain pattern.
Persistence & Privilege
The skill does not request persistent installation privileges (always: false) and declares it will not modify agent config or write files. Autonomous invocation is enabled by default, which is normal; there is no request for elevated or cross-skill configuration changes.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install veo-video-gen - After installation, invoke the skill by name or use
/veo-video-gen - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.7
veo-video-gen 1.0.7
- Added a critical warning to never alter any API response values (such as task_id, result_urls, image_url, upload_url, or fields): always use them exactly as returned from the API to avoid errors or broken links.
- Emphasized storing and passing API response tokens verbatim in all example workflows and instructions for both text-to-video and image-to-video tasks.
- No changes to code or implementation; documentation and usage guidance update only.
v1.0.6
veo-video-gen 1.0.6
- Updated setup instructions to clarify API key provision and verification through standard environment variables or secret managers.
- Now instructs users to provide the Pixwith API key via their runtime environment or hosting platform, not by file editing.
- Enhanced error guidance for cases of missing or invalid API keys.
- No functional changes to workflows or capabilities.
v1.0.5
No user-visible changes in this release.
- Version bump only; no updates or file changes detected.
v1.0.4
- Added a dedicated metadata section with version, publisher, homepage, categories, tags, and environment variable details.
- Explicitly listed required permissions for network access and filesystem usage.
- Declared skill capabilities, such as check_credits, text_to_video, image_to_video, image_upload, task_creation, and task_status_polling.
- No changes to core workflow or functionality.
v1.0.3
- Updated skill setup instructions: the environment variable configuration key is now "veo-video-gen" instead of "pixwith-veo3-1".
- No other changes detected to functionality, workflows, or defaults.
v1.0.2
**Major update: Skill now focuses on generating Veo 3.1 videos through the Pixwith API, supporting text-to-video and image-to-video.**
- Removed non-video documentation and all sample/reference files not relevant to Veo 3.1 video generation.
- Skill now requires only a PIXWITH_API_KEY as environment variable for setup.
- Detailed, streamlined instructions for text-to-video and image-to-video workflows, including API usage and file upload steps.
- Clarifies Fast (quick preview) and Pro (HD with audio) tiers, defaulting to Fast unless user requests otherwise.
- Table-based summaries of pricing, tier differences, parameters, and error handling.
- All previous references, snapshots, and examples for image generation removed to reflect new focus.
v1.0.1
- Minor update: documentation formatting improved for setup guidance.
- Added bullet points to setup instructions for clarity.
- No changes to functionality or tool logic.
v1.0.0
- Initial release of veo-video-gen skill for integration with Pixwith AI via MCP.
- Supports generating images and videos from text or reference images.
- Guides users through first-time Pixwith MCP setup, including API key configuration.
- Handles model selection dynamically and enforces schema-based parameter validation.
- Enables job status polling and credit checks before long-running or costly tasks.
- Provides ready-to-use MCP configuration examples and setup instructions.
Metadata
Frequently Asked Questions
What is (Google) Veo 3 Video Gen?
Generate videos using Pixwith API's Veo 3.1 model. Supports text-to-video and image-to-video (start/end frames). Two tiers: Fast (quick preview) and Pro (HD... It is an AI Agent Skill for Claude Code / OpenClaw, with 342 downloads so far.
How do I install (Google) Veo 3 Video Gen?
Run "/install veo-video-gen" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is (Google) Veo 3 Video Gen free?
Yes, (Google) Veo 3 Video Gen is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does (Google) Veo 3 Video Gen support?
(Google) Veo 3 Video Gen is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created (Google) Veo 3 Video Gen?
It is built and maintained by mango_fish (@tate-kt); the current version is v1.0.7.
More Skills