Send a public video URL directly to a Google Gemini model for analysis. Use when Codex must summarize a video, answer questions about video content, or extra...

This skill appears to do what it says (send a public video URL to Google Gemini and print the text result), but the package metadata omitted important runtime requirements. Before installing or running it: (1) be aware the script will send the video URL and your prompt to Google — do not use it with private/personal videos or sensitive content; (2) supply a GEMINI_API_KEY (or GOOGLE_API_KEY) and prefer a minimally scoped Gemini key rather than a broad GOOGLE_API_KEY; (3) you'll need the 'google-genai' Python package (pip install google-genai); (4) the registry should have declared the required env var and dependency — treat the omission as a sign of sloppy packaging and prefer to inspect or pin the dependency before use. If you need higher assurance, ask the maintainer to update metadata to list required env vars and dependencies and to provide a pinned install spec.

Type: OpenClaw Skill Name: gemini-video-analyze Version: 1.0.0 The skill bundle is benign. It provides a Python script (`scripts/gemini_video_url_analyzer.py`) to analyze public video URLs using the Google Gemini API. The script securely handles API keys from environment variables or command-line arguments and passes user-provided video URLs and prompts to the `google-genai` library for interaction with Google's legitimate API. There is no evidence of data exfiltration beyond the necessary API key for authentication, no malicious execution (e.g., shell injection, arbitrary code execution), no persistence mechanisms, and no prompt injection attempts in the `SKILL.md` or `agents/openai.yaml` that would lead to unauthorized actions by the agent.