Supplier Video Ad Builder

Transforms supplier or CJ source videos into 1080×1920 TikTok/Instagram Reels ads with clean zone detection, Pillow text overlays, CTA card, and trending audio.

This skill appears to do exactly what it claims: local video/image processing using ffmpeg and Pillow. Before running: (1) verify ffmpeg/ffprobe and Python/Pillow are installed; (2) review and use the example config but point to trusted local files (font, logo, source videos, audio); (3) run initially on non-sensitive sample media in an isolated directory to confirm output and avoid accidental overwrites; (4) note there is no network activity or credential use, but the source code and owner are provided without an external homepage — if provenance matters for you, consider asking the publisher for more info or running in a sandbox. Minor operational notes: filenames with unusual characters may break ffmpeg concat quoting, and missing/invalid ffprobe output can cause exceptions; these are implementation robustness issues, not indicators of malicious behavior.

Type: OpenClaw Skill Name: skill-supplier-video-ad Version: 1.0.0 The skill bundle provides a legitimate utility for automating the creation of short-form video ads using ffmpeg and the Pillow library. The Python scripts (detect_clean_zones.py and make_product_ad.py) are well-structured and use subprocess.run with argument lists, which correctly mitigates shell injection risks. The workflow described in SKILL.md is consistent with the code's functionality, and no indicators of data exfiltration, persistence, or malicious prompt injection were identified.