by 安全扫描
OpenClaw
安全
high confidenceThe skill's requirements and runtime instructions are consistent with its stated purpose (generating videos via Google's Veo and Gemini APIs); it asks only for a Google API key and standard CLI/Python tools and doesn't attempt unexpected data exfiltration or unrelated credential access.
评估建议
This skill appears to do what it says: it uploads an image to Google (Gemini/Veo) and saves the returned video under ~/.openclaw/workspace/tibetanProc/. Before installing/using: 1) Confirm you are comfortable providing a GOOGLE_API_KEY (the key will be used client-side to call Google APIs and may incur billing). 2) Avoid uploading sensitive images — they are transmitted to Google's services. 3) Ensure the required Python packages (google-generativeai, requests) are installed in the environment; ...详细分析 ▾
✓ 用途与能力
Name/description (image→video using Veo + Gemini) matches the declared requirements: GOOGLE_API_KEY for Google APIs, python3 and small CLI tools for the provided quickstart. Declared Python packages (google-generativeai, requests) are what the example code uses.
✓ 指令范围
SKILL.md instructions stay within scope: analyze image with Gemini, call Veo predictLongRunning, poll and download video to ~/.openclaw/workspace/tibetanProc/. The script reads only the image file and GOOGLE_API_KEY and writes outputs to the stated workspace. It uploads the image to Google services (expected for Gemini/Veo).
✓ 安装机制
Instruction-only skill (no install spec or code files) — low install risk. It does require certain Python packages but provides no installer; this is an operational note rather than a security contradiction.
✓ 凭证需求
Only GOOGLE_API_KEY is requested, which is proportionate for calling Google generative APIs. No unrelated secrets or system config paths are requested. Be aware an API key can incur billing and will be sent to Google endpoints.
✓ 持久化与权限
Skill does not request persistent 'always' inclusion or elevated platform privileges. It only writes and reads files under the declared ~/.openclaw workspace directory.
安全有层次，运行前请审查代码。
运行时依赖
无特殊依赖
版本
latestv3.0.02026/4/11
image-to-video-gen v3.0.0 — Major update: introduces robust, fully working Veo-3.0 async REST API integration. - Gemini Vision now performs image analysis to generate intelligent motion/scene prompts. - Veo-3.0 video generation uses correct REST payload (`bytesBase64Encoded`), enabling successful async job submission. - Polling loop reliably tracks operation status and downloads generated MP4 upon completion. - All intermediary files and outputs are now consistently saved with timestamped names in `~/.openclaw/workspace/tibetanProc/`. - Sample workflows, payloads, and error handling updated for clarity and recent API requirements.
● 无害
安装命令
点击复制官方npx clawhub@latest install image-to-video-gen
镜像加速npx clawhub@latest install image-to-video-gen --registry https://cn.longxiaskill.com镜像同步中