Anti-AI-Slop Design

Make visual decisions follow the product, audience, task, and brand. An AI-enabled product does not require AI-looking design. Apply these principles independently or alongside a visualization, product-design, image-generation, or implementation skill.

Workflow

1. Inspect the supplied screenshot, mockup, existing design, brief, or relevant project context before judging it. For repository work, inspect the smallest useful set of screens, components, tokens, and product documents.
2. Identify the primary user task, focal information, established brand, supported functionality, and requested scope. Distinguish observed defects from assumptions; without a visible artifact, give guidance rather than pretending to audit a screen.
3. Find the few generic patterns that most weaken meaning, hierarchy, or usability. Explain their concrete effect when a review is requested.
4. Choose one coherent correction based on product context. Preserve functioning information architecture, interactions, exact required wording, and unrelated regions.
5. Match the output to the request: report findings for an audit; return a usable brief for a prompt request; generate an image for a requested raster concept when an image tool is available; edit code only when implementation is requested. Use the relevant tool or skill for execution and verification.
6. Verify the result against the acceptance checks below. Make conservative, reversible choices when routine creative details are missing. Ask only if uncertainty materially changes product meaning, identity, or required content.

Remove generic defaults

Avoid these choices unless an explicit request, established brand, or clear functional purpose justifies them: