Videogen

Things to consider before installing or running this skill: - Metadata mismatch: The SKILL.md clearly requires external API keys (MiniMax Hailuo and optional IMA) and local tooling (python, ffmpeg, node/npm, pip packages), but the registry metadata lists none. Treat the absence of declared env vars/binaries as a red flag — ask the author to declare required credentials and system prerequisites explicitly. - Data exposure: The pipeline will send input content (text, webpage content, possible images) to third‑party services (Hailuo, IMA, Whisper/OpenAI if enabled). Only run it with accounts/keys you trust and that are scoped appropriately; avoid using high‑privilege or production API keys. - Run in a sandbox: Because the package contains many scripts and will execute arbitrary python and shell commands and call ffmpeg, run it in an isolated environment (container, VM) with limited network and filesystem access until you review the code you will execute (notably scripts/v2/*.py and skills/minimax-multimodal/scripts/video/generate_video.py). - Inspect the largest scripts: review run_pipeline.py, tts_harness.py, generate_video.py, url_extractor.py to confirm which external endpoints they call, how they source API keys (env vars vs config files), and what they log or upload. Look for hardcoded endpoints, telemetry, or unexpected uploads. - Verify package sources: npm lockfile entries resolve to a Tencent mirror (http://mirrors.tencentyun.com). If you plan to npm install, prefer using official registries (https://registry.npmjs.org) or verify the integrity hashes before trusting those archives. - Ask the author for clarifications you need: an explicit list of required environment variables and their names, exact endpoints used by generate_video.py, whether any secrets are written to disk or transmitted to other services, and a minimal safe run recipe (e.g., a dry‑run mode that does not call external APIs). If you want, I can enumerate the top files to inspect (run_pipeline.py, url_extractor.py, tts_harness.py, generate_video.py) and summarize any external network calls or uses of environment variables I find inside them.