Agent skill · software engineering · hkuds
quietshrink
Compress macOS screen recordings with zero CPU stress using Apple Silicon's hardware HEVC encoder. Typically reduces file size 70-90% while staying visually lossless. Computer stays silent during encoding.
Why this skill is useful
Adds specific commands and quality presets for efficient video compression using Apple Silicon's hardware capabilities that the AI wouldn't reliably generate on its own.
What it needs
Requires ffmpeg installed locally. About 2k tokens when loaded. Last updated 2026-08-03. 46,739 stars on the source repository.
What this skill does
quietshrink — Agent Skill You have access to cli-anything-quietshrink, a CLI for compressing video files on macOS Apple Silicon. It uses the Media Engine (hardware HEVC encoder), not the CPU, so encoding is fast and silent. When to use it User wants to compress a screen recording, screencast, or any .mov/.mp4 file File is too large to share (chat, email, GitHub) Need smaller files but cannot tolerate visible quality loss Apple Silicon Mac (M1/M2/M3/M4) — best results Don't use this on non-screen content (camera footage, vlogs) — savings will be much smaller because there are no duplicate frames to drop. Commands All commands accept --json for machine-readable output. Quality presets Preset q Typical reduction SSIM Use case -------- --- ------------------- ------ ---------- tiny 50 ~90% ~0.95 Chat/email — small artifacts OK balanced 55 ~88% ~0.99 Docs/sharing — high quality transparent (default) 60 ~87% ~0.99+ Anything important — visually lossless pristine 70 ~84% ~0.997 Archival — near-source JSON output schema compress returns: probe returns: Decision flow for agents Before processing, run doctor to verify environment. For unfamiliar files, run probe to understand resolution/codec/duration. Errors and recovery ffmpeg not found → brew install ffmpeg hevcvideotoolbox not available → brew reinstall ffmpeg compressionfailed → check input file isn't corrupted; try --verbose mode to see ffmpeg errors Source Main repo: https://github.com/achiya-automation/quietshrink Bash CLI: quietshrink command (via the install script) Why this approach works: see WHY.md
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills hkuds/skills--036d4b