Veo 3 Video Gen (Gemini API)

Generate and stitch short videos via Google Veo 3.x using the Gemini API (google-genai). Use when you need to create video clips from prompts (ads, UGC-style clips, product demos) and want a reproducible CLI workflow (generate, poll, download MP4, optionally stitch multiple segments).

This skill appears to implement Veo/Gemini video generation as advertised, but the package metadata is inconsistent and the source is 'unknown.' Before installing or running it: (1) Treat GEMINI_API_KEY as required — only provide an API key you trust and that has appropriate billing/quota limits. (2) Be prepared to install python deps (google-genai, pillow) and ensure ffmpeg is on PATH if you plan to concatenate segments. (3) Review the script contents locally (it is included) — there are no hidden remote endpoints, but it uses subprocess/ffmpeg and prints MEDIA: lines which some hosts may auto-attach/upload outputs. (4) Run in an isolated environment or test account if you are unsure about billing or data exfiltration. (5) Ask the publisher (or avoid) if you need the package metadata corrected — the omission of required env vars/install instructions is a red flag for sloppy or incomplete packaging.

Type: OpenClaw Skill Name: veo3-video-gen Version: 0.1.0 The skill is classified as benign. It uses `subprocess.run` in `scripts/generate_video.py` to execute `ffmpeg` for video concatenation and last-frame extraction, which are high-risk capabilities. However, `ffmpeg` is explicitly listed as a requirement in `SKILL.md` and its use is directly aligned with the skill's stated purpose of generating and stitching videos. The commands passed to `ffmpeg` are constructed programmatically from controlled inputs, and there is no evidence of data exfiltration, malicious execution, persistence, or prompt injection attempts.