Screenshot Annotator
SkillCommerce & financeLets your agent capture screenshots, add annotations, and share them with AI-generated descriptions.
Available today. Use it from your connected AI after setup.
No other account needed.
Add ahel to your AI once: Claude, ChatGPT, Cursor, Claude Code or Codex. Then ask it to use this.
Then ask your AI: use the Screenshot Annotator skill
About this skill
Capture, annotate, and share screenshots with AI-powered descriptions using Peekaboo.
What this skill tells your AI
The instructions your AI receives, as published by jiayaoqijia/cryptoskill in skills/ai-crypto/ianalloway-screenshot-annotator/SKILL.md and read by ahel’s review.
Capture screenshots with automatic UI element annotation and AI-powered descriptions. Built on top of Peekaboo for macOS UI automation.
Features
- Capture full screen, specific windows, or regions
- Automatically annotate UI elements with clickable IDs
- Generate AI descriptions of what's on screen
- Save annotated screenshots with metadata
- Compare before/after screenshots
- Create step-by-step visual guides
Requirements
- macOS with Screen Recording permissions
- Peekaboo CLI installed (
brew install openclaw/tap/peekaboo)
Quick Start
Capture and annotate current screen
peekaboo see --annotate --path /tmp/annotated-screen.png
Capture specific app window
peekaboo see --app "Safari" --annotate --path /tmp/safari-annotated.png
Capture with AI analysis
peekaboo see --annotate --analyze "Describe the main UI elements and their purpose" --path /tmp/analyzed.png
Common Workflows
Document a bug
# 1. Capture the problematic state
peekaboo see --app "MyApp" --annotate --path /tmp/bug-screenshot.png
# 2. Get element IDs for the bug report
peekaboo list windows --app "MyApp" --json
# 3. Describe what's wrong
peekaboo see --app "MyApp" --analyze "Identify any UI issues or anomalies"
Create a tutorial
# Step 1: Capture initial state
peekaboo see --app "Finder" --annotate --path /tmp/step1.png
# Step 2: Click on target element
peekaboo click --on B3 --app "Finder"
# Step 3: Capture result
peekaboo see --app "Finder" --annotate --path /tmp/step2.png
Compare UI states
# Before
peekaboo see --app "Settings" --path /tmp/before.png
# Make changes...
# After
peekaboo see --app "Settings" --path /tmp/after.png
# Analyze differences
peekaboo see --app "Settings" --analyze "Compare this to the previous state and describe changes"
Advanced Usage
Capture specific region
peekaboo image --mode region --region 100,100,800,600 --path /tmp/region.png
Capture with retina resolution
peekaboo image --mode screen --retina --path /tmp/retina-screen.png
Capture menu bar
peekaboo menubar list --json
peekaboo see --mode menubar --annotate --path /tmp/menubar.png
Batch capture all windows
for app in "Safari" "Finder" "Terminal"; do
peekaboo see --app "$app" --annotate --path "/tmp/${app}-annotated.png"
done
Output Formats
Screenshots are saved as PNG by default. Use --format jpg for JPEG output.
The --json flag outputs metadata including:
- Window dimensions and position
- Annotated element IDs and locations
- Timestamp and app info
Tips
- Always use
peekaboo see --annotatebefore clicking to identify element IDs - Use
--analyzewith specific prompts for better AI descriptions - Combine with
peekaboo clickandpeekaboo typefor automated workflows - Check permissions with
peekaboo permissionsif captures fail
Author
Created by Ian Alloway - Data Scientist specializing in AI/ML.
License
MIT License
Signals
- GitHub stars
- 77
- Forks
- 16
- Last commit
- Sep 2026
ahel review
K1binfo
installs-packagesK1info
remote-installer-piped-to-shell (in TRUST.auto.yaml)
Automated review, not a security audit. Ruleset v1+k2.
Advanced
- Item type
- skill
- Key
screenshot-annotator- Source
- github.com/jiayaoqijia/cryptoskill