Claude Skill MartBrowse skillsQuick linesLearn by videoTerminal guideWhat is a Skill?
 
← Back to list

video-scrub — Total Video Metadata Stripper

Rebuilds a video with picture and sound only — no GPS, device IDs, chapters, telemetry or hidden encoder fingerprints — then proves it with a byte-level scan.

Image & VideoIntermediate★ 805⑂ 99AI score 9/10Last updated: Sep 21, 2026

What it does

  • Whitelist remux, not blacklist deletion: instead of enumerating fields to delete, it carries over only one video stream and one audio stream. Vendor-private udta boxes, GoPro/DJI GPS telemetry tracks, and chapters have no path into the output at all.
  • Kills the invisible layers: blocks five hiding places ffprobe cannot show you — the x264 parameter string in H.264 SEI, Lavc... in the AAC bitstream DSE, avc1 compressorname, and more.
  • Two modes: copy (video copied byte-for-byte, audio re-encoded — 1.3s for a 53s clip) and encode (full re-encode, also removes bitstream-domain marks).
  • Four profiles: obs (default), quicktime, screencapture, and bare (writes nothing — strongest privacy).
  • 12 verification gates: every metadata string from the source becomes a "needle" scanned byte-level in the output; a single hit fails the run. Needles must be ≥5 bytes and generic strings are excluded to avoid false positives.
  • HDR & caption safety: SEI also holds HDR mastering metadata and CEA-608/708 captions, so the SEI cut is applied only when an identity string is actually found, and never on HDR sources.

Who it's for

  • Creators who want location and device model gone before posting footage
  • GoPro / DJI / phone users worried about GPS telemetry tracks that tag-strippers miss
  • Anyone distributing screen recordings without leaving tooling fingerprints
  • Journalists and researchers handling video that must stay anonymous

Examples

  1. Audit first: "Show me what metadata this clip carries" → inspect lists container tags, per-stream tags and side data, chapters, plus the bitstream identity strings ffprobe hides.
  2. Everyday one-shot: "Strip all metadata from this mp4 and verify it" → run scrubs and then executes all 12 gates.
  3. Maximum privacy: "Full re-encode, declare nothing" → --mode encode --profile bare.

Note: this is not a watermark remover. Robust pixel watermarks (SynthID, forensic marks) survive re-encoding by design. Only mp4/mov containers, one file at a time, no batch directory scans.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/eternityspring/reelbench-skills/HEAD/skills/video-scrub/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the skills/video-scrub folder from the GitHub repo eternityspring/reelbench-skills into my ~/.claude/skills/video-scrub/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/eternityspring/reelbench-skills.git && mkdir -p ~/.claude/skills && cp -r reelbench-skills/skills/video-scrub ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Install prerequisites: you need node (>= 18) and ffmpeg/ffprobe. macOS: brew install node ffmpeg; Ubuntu/Debian: sudo apt install nodejs ffmpeg.
  2. Verify them: run node -v and ffmpeg -version and confirm both print versions.
  3. Clone the repo: git clone https://github.com/eternityspring/reelbench-skills.git
  4. Copy the skill in: mkdir -p ~/.claude/skills && cp -r reelbench-skills/skills/video-scrub ~/.claude/skills/
  5. Run the self-test: node ~/.claude/skills/video-scrub/scripts/selftest.mjs — 117 assertions, no ffmpeg or real files needed.
  6. Restart Claude Code, then ask "strip metadata from this video" or "scrub video" to trigger it.
  7. On your first run, use inspect before scrubbing and always finish with verify (or just use run) to prove the output is clean.
View source on GitHub ↗License: Apache-2.0