Screen Reviewer for Openclaw

Screen Reviewer is an automated monitoring tool that captures screenshots, extracts text via OCR, and uses AI to generate structured daily productivity reports.

drpris
v1.0.0
Mar 30, 2026
0
673
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install screen-reviewer

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install screen-reviewer using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is Screen Reviewer?

Screen Reviewer is a sophisticated activity logging and ROI analysis tool designed for the Openclaw Skills ecosystem. It functions by silently monitoring computer activity through periodic screenshots and advanced text extraction, transforming raw screen data into meaningful insights.

By leveraging AI models like GPT-4o or Claude, this skill categorizes your workflows, identifies time sinks, and provides a structured daily review. It is built with privacy in mind, allowing for app blacklisting and smart detection to ensure only relevant, changed frames are processed and stored.

Screen Reviewer Use Cases

  • Automatically logging daily work activities for professional time tracking and billing.
  • Generating daily productivity reviews and ROI analysis to optimize focus and deep work.
  • Maintaining a searchable history of on-screen text and window titles for personal knowledge management.
  • Monitoring specific app usage patterns to identify and eliminate digital distractions.

How Screen Reviewer Works

  1. The capture loop runs every 5 seconds, taking a screenshot and comparing it to the previous frame to detect significant pixel changes.
  2. The system retrieves the active window title and application name while performing OCR to extract visible text from the screen.
  3. Activity data is saved into a structured JSONL log file with associated JPEG screenshots stored in a local directory.
  4. On-demand or at a scheduled time (e.g., 8 AM), the reporting engine aggregates these logs and classifies apps based on value tiers.
  5. An AI provider processes the aggregated data to generate a comprehensive Markdown report featuring productivity scores and activity summaries.
  6. Automated cleanup tasks delete screenshots older than 3 days to manage disk space effectively.

Screen Reviewer Setup

To install Screen Reviewer for Openclaw Skills, run the following command from the root directory:

bash install.sh

After installation, you must grant Screen Recording and Accessibility permissions in System Settings > Privacy & Security. To start the capture service, use the service manager:

~/.screen-reviewer/venv/bin/python scripts/service_manager.py start

Screen Reviewer Data Schema & Taxonomy

The skill organizes its data within the ~/.screen-reviewer/ directory using the following structure:

File/Folder Purpose
/screenshots/ Periodic JPEG captures organized by date folders.
/logs/YYYY-MM-DD.jsonl Structured logs containing timestamps, app names, window titles, and OCR text.
/reports/ AI-generated Markdown reports (e.g., 2026-03-22-review.md).
config.yaml User configuration for intervals, blacklists, and AI model settings.
daemon_stderr.log Error logs for troubleshooting the background service.

Screen Reviewer Advanced Features

  • Smart Frame Detection: Automatically skips capturing and processing if the screen content hasn't changed beyond a 5% threshold.
  • Privacy Blacklisting: Support for excluding specific sensitive applications like 1Password from the capture loop.
  • Multi-Provider AI Support: Flexible integration with OpenAI, Claude, or local Ollama instances for report generation.
  • Automated ROI Mapping: Customizable app-to-value mapping to calculate the return on investment for your daily time usage.
  • macOS Auto-start: Built-in commands to install or uninstall the service as a launchd agent for persistent monitoring.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*