- Home
- Skills
- Testing & Debugging
- Catch Visual Changes in Desktop Apps
More screenshots
Works with the AI Agents you already use
Catch Visual Changes in Desktop Apps
Save a screenshot of how your app should look, and get told the moment a change makes it look different.
$9
Catch Visual Changes in Desktop Apps
Example session with this skill installed
Use the pixel-diff skill to check whether my latest desktop app change broke the UI visually: compare the current screenshots against the baselines and tell me if anything actually regressed.
- Read your context and instructions
- Compiled the catch visual changes
Visual regression check: 1 real difference found
Baselines: 8 screens, pinned at 1440x900, created with --update-baselines after visual review.
Result: 7 screens identical, 1 screen differs.
Failing screen: Settings — sidebar
- Diff pixels: 2.1% (max-diff-ratio 1.0% exceeded)
- Red diff image shows the sidebar shifted 12px left after the layout change.
Verdict: Real regression — the sidebar spacing changed unintentionally. Not an accepted visual change, so I did NOT update the baseline.
Fix: Restored the sidebar padding, re-ran the comparison — all 8 screens now match their baselines.
CI note: Baselines are only blessed after viewing them, and runs in CI use the same pinned resolution as the baselines, so results are reproducible.
Connects securely to your tools. The creator never sees your data.
About this skill
Refactors and theme changes break layouts quietly. The code still runs; the screen is wrong, and nobody notices for weeks. This skill teaches your agent screenshot comparison for desktop apps: accept a baseline picture of a view, then fail the test whenever a later screenshot differs from it by more than a tolerance you set.
Inside the package: the skill file covering the accept-a-baseline workflow, a tuning guide for the per-pixel tolerance and how much difference to allow, a filled example test, and a config example. It runs on top of deskcheck-ai, which is free and MIT-licensed; this package is the method and the judgment.
A run goes like this: the agent screenshots a view, and the first run fails on purpose because no baseline exists yet. You look at the screenshot yourself and accept it as the baseline only if the screen is actually right. From then on, every run compares. A failure produces a diff image with the changed areas marked in red. You decide what it means: real breakage (fix the app, keep the baseline) or a change you meant to make (accept a new baseline, after viewing it). The skill is strict about never raising the tolerance just to make a failure go away.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 6 days ago
Requires