agentsclimarketplace

Design mode

Skill Adit-Jain-srm/skill-forge/skills/design-mode

18 AI agent skills that catch what your agent misses: hidden code debt, false completions, missing error handling, context amnesia, imprecise UI edits. Self-improving with TF-IDF routing + SkillOpt loop. Cursor, Claude Code, Codex, Copilot. Install: npx skills@latest add Adit-Jain-srm/skill-forge

Install
npx -y skills add Adit-Jain-srm/skill-forge --skill design-mode

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Use when making UI/frontend changes in Cursor's Design Mode, when the user selects elements visually, draws annotations, or uses voice commands to describe changes. Also use when editing components from the integrated browser, when multi-selecting elements to match styles, when making iterative visual refinements, or when the agent receives visual context (screenshots, element selections, DOM references) alongside a change request.

SKILL.md

4.8 KB, as published. Nobody here has run it

Overview

Cursor Design Mode lets users point, click, draw, and speak to direct UI changes — giving you spatial context, component identity, and visual relationships. This skill teaches you to use that context effectively: make targeted, precise edits instead of broad rewrites.

When to Use

  • User selected element(s) in the integrated browser
  • User drew an annotation highlighting a region
  • User gave voice instruction about a visual change
  • You received DOM context, layout data, or element identity from Design Mode
  • Task involves CSS, layout, spacing, color, typography, or component styling
  • User says "make this look like X" or "fix this button" with visual context

Process

When Design Mode provides visual context:

1. IDENTIFY — What exact element(s) were selected? Which component file owns them?
2. SCOPE — Change ONLY what was pointed at. Don't refactor surrounding code.
3. PRESERVE — Keep existing styles, classes, and behavior. Only modify what was asked.
4. TARGETED — Make the minimal CSS/JSX change that achieves the visual result.
5. VERIFY — After edit, confirm the change renders correctly (hot-reload shows result).

Key Principles

Precision Over Ambition

Design Mode gives you a SPECIFIC element. Don't:

  • Refactor the entire component when only one style was requested
  • Change the component's API when only its appearance was pointed at
  • Touch logic when only visuals were selected

DO:

  • Find the exact file and line for the selected element
  • Make the minimal edit that achieves the described visual change
  • Preserve all existing behavior, event handlers, accessibility

Parallel Edits

Users can send multiple Design Mode edits simultaneously. Each edit is independent:

  • Don't wait for one to complete before processing another
  • Each edit targets its own element — no cross-contamination
  • If two edits conflict (same element, different requests), handle the most recent

Voice + Visual = High Confidence

When user provides BOTH voice description AND visual selection:

  • The visual selection tells you WHAT to change
  • The voice tells you HOW to change it
  • Trust the combination — this is higher confidence than text alone

Workflow Patterns

Single Element Edit

User: [clicks element] "Make this text larger and bolder"
Agent:
1. Identify: The selected h2 in src/components/Hero.tsx line 23
2. Change: font-size + font-weight on that specific element
3. Scope: Only the selected h2, nothing else

Multi-Element Style Matching

User: [multi-selects 3 cards] "Make these all the same height"
Agent:
1. Identify: The 3 card components — check if they share a class/component
2. Change: Add consistent height/min-height to the shared style
3. Verify: All 3 cards now match visually

Region Drawing

User: [draws rectangle around a section] "Add more padding here"
Agent:
1. Identify: The container element within the drawn region
2. Change: Increase padding on that container
3. Scope: Only the highlighted region, not global spacing

Common Mistakes

  • Rewriting an entire component when user pointed at one button
  • Changing Tailwind config when user just wants one element's color adjusted
  • Breaking responsive behavior by adding fixed widths to match a desktop selection
  • Ignoring the specificity of the selection (user clicked ONE card, you changed ALL cards)
  • Removing existing accessibility attributes (aria-*, role, tabindex) during visual edits

Stack Guidance

Design Mode works best with:

  • Tailwind CSS — utility classes make targeted edits easy (add/change one class)
  • CSS Modules — scoped by default, safe to edit without side effects
  • Inline styles — only for truly one-off overrides the user explicitly requested
  • Component props — when the visual change maps to an existing variant/size/color prop

Avoid:

  • Global CSS changes for element-specific requests
  • !important unless overriding a third-party library with no other option
  • Modifying shared design tokens for a single-element request

Integration with Composer 2.5

Design Mode pairs with Composer 2.5 for optimal speed on UI tasks. When processing Design Mode edits:

  • Prefer fast, targeted changes over architectural refactors
  • Hot-reload feedback loop: edit → save → see result → iterate
  • If the first edit doesn't look right, make a second targeted adjustment rather than reverting and trying a different approach

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.