screen-vision — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited screen-vision (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
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.
Use this skill when the user wants Claude to see what's on their screen, watch their screen over time, analyze videos, or extract text.
capture_screen(delay_seconds, monitor, scale) — Full screen capture. Use delay for window switching.capture_region(x, y, width, height, scale) — Capture a specific rectangular area.capture_window(window_title, scale) — Capture a window by title (e.g., "Chrome", "Terminal").list_monitors() — List all displays with resolutions and positions.get_active_context() — Lightweight window/cursor/monitor info without a screenshot.read_screen_text(region) — OCR on screen or region. Optional region as "x,y,width,height".understand_screen(prompt) — AI-powered screen analysis via GenAI Gateway.analyze_image(file_path, prompt) — Analyze any image file (e.g., AirDropped photo).watch_screen(duration_seconds, interval_seconds, include_audio, max_frames) — Watch screen with frame sampling + optional audio transcription via Whisper.analyze_video(file_path, start_time, end_time, max_frames) — Extract keyframes from local video files.Screen Vision includes security controls:
~/.screen-vision/audit.logThese are not required — tools gracefully degrade without them:
brew install tesseractbrew install ffmpegUser: "What's on my screen?"
Claude: [calls capture_screen()]
User: "Watch my screen while I demo this feature"
Claude: [calls watch_screen(duration_seconds=60, include_audio=true)]
User: "Analyze the video at ~/Downloads/demo.mp4"
Claude: [calls analyze_video(file_path="~/Downloads/demo.mp4")]
User: "Read the text on my screen"
Claude: [calls read_screen_text()]
User: "What window am I in?"
Claude: [calls get_active_context()]capture_screen with delay for window switching: "take a screenshot in 5 seconds"watch_screen with audio for demos where user is talkingcapture_region for specific UI elementsread_screen_text when user wants text extraction only (faster than full capture)get_active_context for lightweight queries (no image processing)~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.