Wjs Reframing Video

PassAudited by ClawScan on May 13, 2026.

Overview

The skill appears to do the advertised local video reframing, with noteworthy but purpose-aligned use of ffmpeg, Python media libraries, and a first-run MediaPipe model download.

Before using, be comfortable installing the documented media dependencies and allowing a first-run MediaPipe model download. Use it on videos you choose, review the output path and .crop.json sidecar, and note that the supplied artifacts do not show credential access, uploads, or background behavior.

Findings (2)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

First use may require network access and depends on the integrity and availability of the external model source.

Why it was flagged

The helper downloads an external MediaPipe model on first use and caches it locally. This is disclosed and purpose-aligned, but the artifacts do not show a checksum or pinned integrity verification for the downloaded model.

Skill content
MODEL_URL = ("https://storage.googleapis.com/mediapipe-models/face_landmarker/" ... "face_landmarker.task") ... urllib.request.urlretrieve(MODEL_URL, MODEL_PATH)
Recommendation

Use in an environment where downloading the MediaPipe model is acceptable; for stricter environments, pre-download and verify a trusted model copy.

What this means

The skill will run local media-processing binaries against videos you select, which can consume CPU/disk and relies on the installed ffmpeg/ffprobe tools.

Why it was flagged

The script invokes local ffprobe and ffmpeg commands to inspect and sample the user-provided video. This command execution is central to the stated video-processing purpose and uses argument lists rather than shell interpolation.

Skill content
subprocess.run(["ffprobe", ... str(path)], check=True, capture_output=True, text=True) ... subprocess.run(["ffmpeg", "-nostdin", "-y", "-i", str(input_path), "-vf", f"fps={sample_fps}", ...], check=True
Recommendation

Only run it on intended video files and ensure ffmpeg/ffprobe and Python media packages are installed from trusted sources.