visual-edit-precision

Makes minimal targeted UI edits based on visual context like element selections and screenshots.

39.3k|4.2k|Updated Jul 24, 2025
One-click install
npx skills add https://github.com/wshobson/agents --skill visual-edit-precision
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: visual-edit-precision
Source: https://github.com/wshobson/agents/tree/main/plugins/skill-forge-essentials/skills/visual-edit-precision
Command: npx skills add https://github.com/wshobson/agents --skill visual-edit-precision

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

When users point at a specific element in a screenshot or annotation and ask for a change, AI agents often overreach—refactoring entire components, breaking accessibility attributes, or altering unrelated styles. This Skill enforces disciplined, minimal edits that change exactly what was indicated and nothing else.

Core Features & Use Cases

  • Element-Scoped Editing: Identifies the exact component file and line owning the selected element, then modifies only that target.
  • Preservation Guarantees: Keeps existing styles, event handlers, responsive behavior, and accessibility attributes (aria-*, role, tabindex) intact during visual changes.
  • Conflict Handling: Processes multiple visual edits independently and resolves conflicts by applying the most recent request for the same element.
  • Use Case: A user circles one card in a dashboard screenshot and asks to change its background color. The Skill locates that specific card's CSS, adjusts only its color, and leaves all other cards and layout untouched.

Quick Start

Change the color of the button I highlighted in this screenshot without touching anything else in the component.

Frequently Asked Questions about visual-edit-precision

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I make UI changes based on a screenshot annotation?

Identify the exact element indicated in the visual context, locate the component file that owns it, and make the minimal CSS or markup change needed. Preserve all surrounding styles, event handlers, and accessibility attributes while editing only what was pointed at.

How to edit one element's style without affecting similar components?

Target the specific element using its unique selector or scoped class rather than shared classes. Avoid changing global config or shared stylesheets when only one instance, such as a single card among many, needs adjustment.

What are common mistakes when editing UI from visual selections?

Common mistakes include rewriting entire components when one button was selected, adding fixed widths that break responsive behavior, and removing accessibility attributes like aria-* or role during visual edits. Change only what was indicated.

How are conflicting edits to the same element handled?

Each visual edit targets its own element independently and can be processed in parallel. When two requests conflict on the same element, the most recent request takes precedence.

When should visual context guide frontend edits instead of text instructions?

Use visual context when the user selects elements, draws annotations, or provides screenshots alongside change requests. It supplements text instructions with spatial data like element identity and DOM references for precise targeting.