Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
macos-screen-recorder (sck-record)
macos-screen-recorder (sck-record) is a macOS screen recorder that captures the screen with system audio via ScreenCaptureKit — CLI, no driver, no sudo — plus a Claude Code skill.
*Record your macOS screen with the sound your Mac is playing — headless, from the CLI — with no virtual audio driver, no BlackHole, no Loopback, no sudo. One Swift file built on [ScreenCaptureKit](https://developer.apple.com/documentation/screencapturekit), Apple's native capture framework.*
sck-record usage: a terminal showing the one-line build (swiftc -O sck-record.swift), a 5-second capture writing demo.mp4 at 1920×1200 H.264 / AAC 48kHz, and an ffprobe confirming both a video stream and an audio stream — "← system audio present" — with a footer noting Screen Recording permission only, no driver install, no sudo
The whole gap, closed: `ffprobe` confirms the muxed file has both a video stream and a system-audio stream — captured with only the standard Screen Recording permission, no driver, no sudo.
🤔 Why
macOS's built-in screencapture -v records video only — its -g flag captures the microphone, not what your speakers are playing. The usual fix is installing a virtual audio device (BlackHole, Loopback, Soundflower) and rerouting your output through it, which changes what you hear and may need a security-approval reboot. ScreenCaptureKit makes that unnecessary: since macOS 13 it can tap system audio directly under nothing but the standard Screen Recording grant. This is the smallest possible wrapper around that capability — one file, one binary, one command.
The one true differentiator: system audio from a headless CLI with zero install. This is not a "better than OBS / Screen Studio" tool — it fills exactly the gap those don't: scriptable capture-with-sound, no driver.
✨ What it does
- 🔊 Main display + system audio — every app sound, browser tab,
sayvoice, and music track your Mac outputs, muxed with the screen. What you hear is untouched while recording. - 🎞️ Standard H.264 / AAC `.mp4` you can share immediately (8 Mbps video, 48 kHz stereo 160 kbps AAC).
- 🖥️ Retina capture, capped at 2560 px wide to keep files sane, up to 30 fps; cursor included (or
--no-cursor). - 🪶 No daemons, no kernel extensions, no audio rerouting — ~90 lines of Swift around
SCStream+AVAssetWriter.
📦 Install (Claude Code plugin)
It ships as a skill — an AI agent (Claude Code) can drive it for you:
/plugin marketplace add connerkward/ckw-skills
/plugin install macos-screen-recorder@connerkwardStandalone (this repo only):
/plugin marketplace add connerkward/macos-screen-recorder-system-audio
/plugin install macos-screen-recorder@macos-screen-recorderSee [docs/agents.md](docs/agents.md) for the agent-driven workflow.
🛠️ Run it by hand (CLI)
Requirements: macOS 13 (Ventura)+ (the SCStreamConfiguration.capturesAudio API), and Xcode Command Line Tools (xcode-select --install).
swiftc -O sck-record.swift -o sck-record # build once
./sck-record demo.mp4 5 # or skip the build: swift sck-record.swift demo.mp4 5sck-record [output.mp4] [seconds] [--no-cursor]| Argument | Default | Description |
|---|---|---|
output.mp4 | ~/Desktop/sck.mp4 | Output file path. Overwritten if it exists. |
seconds | 20 | Duration in seconds (fractional allowed). |
--no-cursor | off | Omit the system cursor (for tools that overlay a smoothed synthetic one). |
The tool prints the output path and stops automatically after the requested duration.
🔐 Permissions
First run triggers macOS's Screen Recording permission prompt for the launching app (your terminal). Grant it under System Settings → Privacy & Security → Screen & System Audio Recording, then re-run. Same permission screencapture uses — no separate audio permission, because ScreenCaptureKit delivers system audio under the screen-recording grant. The microphone is never touched.
⚖️ How it compares
| Approach | System audio | Setup | Notes |
|---|---|---|---|
| sck-record | ✅ native | none | One binary, one permission. |
screencapture -v (built-in) | ❌ | none | -g records the mic, not system output. |
| BlackHole / Loopback / Soundflower | ✅ via reroute | install driver, multi-output device, switch output | Rerouting changes what you hear; installs may need security approval. |
| OBS | ✅ via plugin/driver | install app, configure scenes + audio | Powerful, but heavyweight for "give me an mp4 of the next 10 seconds". |
| QuickTime screen recording | ❌ (mic only) | none | Needs a virtual driver for system audio. |
🔬 How it works
SCStream is configured with capturesAudio = true, yielding system-audio sample buffers alongside the screen frames. Both are muxed in real time into an .mp4 by AVAssetWriter (H.264 video + AAC audio). They stay in sync because both inputs share the stream's presentation timestamps; writing starts on the first complete video frame. Read sck-record.swift — it's ~90 lines.
🚧 Limitations
- Main display only — no window or region selection.
- Fixed duration set up front; no interactive stop key (
Ctrl-Caborts without finalizing). - Output codec/bitrate hardcoded (H.264 8 Mbps / AAC 160 kbps) — it's one dictionary in the source to change to HEVC or other rates.
🗂️ Part of ckw-skills
The capture primitive behind the record→polish workflow: the screencast skill calls this, and screenstudio-alt does the post-production (auto-zoom, idle speed-up, vertical export).
License
MIT © Conner K Ward
🧭 [ckw-skills](https://github.com/connerkward/ckw-skills) — part of Conner K. Ward's collection of Claude Code skills & MCP servers.