# How the Hallmark Slop-Test Works: A 58-Gate Anti-AI-Slop Checklist

> Discover how the Hallmark slop-test works. This 58-gate checklist ensures generated pages meet strict anti-AI design rules, automating revisions for any detected slop.

- Repository: [Hassan El Mghari/hallmark](https://github.com/Nutlope/hallmark)
- Tags: how-to-guide
- Published: 2026-07-13

---

**The Hallmark slop-test is a deterministic 58-gate checklist that runs after every build to guarantee generated pages obey strict anti-AI-sloppiness design rules, forcing automatic revision if any gate detects "slop."**

The Hallmark slop-test serves as the mandatory quality gate in the Nutlope/hallmark repository, ensuring every generated landing page meets rigorous design standards before deployment. This deterministic validation system operates as Step 7 of the build pipeline, checking 58 specific criteria to eliminate common AI-generated design flaws ranging from visual inconsistencies to fabricated content.

## Pre-Emit Self-Critique: The Six-Axis Scoring System

Before emitting any HTML or CSS, Hallmark scores the design on six axes: **Philosophy**, **Hierarchy**, **Execution**, **Specificity**, **Restraint**, and **Variety**. Each axis receives a score from 1 to 5. Any axis scoring below 3 triggers an automatic revision pass, preventing low-quality designs from reaching the validation stage. According to [`skills/hallmark/references/slop-test.md`](https://github.com/Nutlope/hallmark/blob/main/skills/hallmark/references/slop-test.md), these scores are recorded in a stamp comment at the top of every output file.

```css
/* Hallmark · pre‑emit critique: P5 H4 E5 S4 R5 V5 */

```

## The 58-Gate Validation Pipeline

After the page build completes, Hallmark executes the slop-test by reading the canonical gate list from [`skills/hallmark/references/slop-test.md`](https://github.com/Nutlope/hallmark/blob/main/skills/hallmark/references/slop-test.md). The system answers "yes" or "no" to each of the 58 gates. All gates must return **no**—any "yes" result marks the output as *slop* and forces a corrective revision before the build can ship.

The test covers eight enforcement categories:
- Visual design rules
- Structural integrity
- Micro-interaction standards
- Variety requirements
- Implementation quality
- Hero section enrichment
- Content diversification
- Layout safety

### Genre-Scoped Overrides

While some gates are universal, others apply only to specific genres: **editorial**, **atmospheric**, **modern-minimal**, or **playful**. The slop-test file contains inline annotations indicating which gates are relaxed for each genre. For example, gradient text is universally forbidden, but atmospheric pages may retain radial-gradient backgrounds where other genres cannot.

## Integration into the Build Workflow

According to [`skills/hallmark/SKILL.md`](https://github.com/Nutlope/hallmark/blob/main/skills/hallmark/SKILL.md), the slop-test runs as Step 7 of the workflow, immediately following the build phase (Step 6). Results are written into the preview block as either `Slop test: 58/58 ✓` or a list of failing gate numbers. If any gate fails, the builder returns to the relevant step, corrects the issue, and re-runs the test until achieving a clean pass.

## Critical Enforcement Points

The slop-test checks concrete anti-patterns defined across multiple reference files. Key restrictions include:

**Gate 37: Font Family Limits**
No more than three font families may appear in the CSS. The implementation extracts all `font-family` declarations and validates the unique count against this limit.

```javascript
function checkFontFamilies(css) {
  const families = new Set();
  css.match(/font-family:\s*([^;]+);/g).forEach(decl => {
    families.add(decl.split(':')[1].trim());
  });
  return families.size > 3; // true = slop detected
}

```

**Gate 10: Transition All Restrictions**
The use of `transition-all` declarations is prohibited to prevent performance-anti-patterns and unintended animation side effects.

**Gate 34: Horizontal Scroll Prevention**
No horizontal scroll may occur at any breakpoint, ensuring responsive layouts remain constrained to viewport width.

**Gates 8 and 32: Variety Enforcement**
These gates reference the macro-structure stamp stored in [`.hallmark/log.json`](https://github.com/Nutlope/hallmark/blob/main/.hallmark/log.json) to enforce variety across builds, preventing repetitive layout patterns.

**Gates 46 and 54: Content Authenticity**
Referencing specific anti-patterns from [`skills/hallmark/references/anti-patterns.md`](https://github.com/Nutlope/hallmark/blob/main/skills/hallmark/references/anti-patterns.md), these gates reject invented metrics and placeholder names. The test ensures all testimonials, statistics, and company names represent real, verifiable content rather than AI hallucinations.

## Summary

- The Hallmark slop-test is a **deterministic 58-gate checklist** that runs automatically after every build at Step 7 of the workflow.
- **Pre-emit scoring** on six axes (Philosophy, Hierarchy, Execution, Specificity, Restraint, Variety) catches low-quality designs before they reach validation.
- **Genre-scoped overrides** allow context-specific relaxations while maintaining universal prohibitions against AI slop.
- The test enforces hard limits on **font families**, **CSS transitions**, **scroll behavior**, and **content authenticity**.
- Implementation resides in [`skills/hallmark/references/slop-test.md`](https://github.com/Nutlope/hallmark/blob/main/skills/hallmark/references/slop-test.md) and integrates into the Hallmark workflow defined in [`SKILL.md`](https://github.com/Nutlope/hallmark/blob/main/SKILL.md).

## Frequently Asked Questions

### What is the Hallmark slop-test?

The Hallmark slop-test is a deterministic quality gate that evaluates generated web pages against 58 specific anti-sloppiness criteria. It ensures every output adheres to strict design standards by automatically flagging and forcing correction of common AI-generated design flaws before deployment.

### How does the pre-emit critique work?

Before emitting any HTML or CSS, Hallmark scores the design on six axes: Philosophy, Hierarchy, Execution, Specificity, Restraint, and Variety. Any axis scoring below 3 out of 5 triggers an automatic revision pass, and the scores are recorded in a stamp comment at the top of the generated file.

### What happens if a gate fails during the slop-test?

If any of the 58 gates returns "yes" (indicating slop detected), the build is marked as failed and cannot ship. The system returns to the relevant workflow step, applies corrections, and re-runs the test until all 58 gates pass, displaying `Slop test: 58/58 ✓` in the preview block.

### Are all 58 gates applied to every genre?

No, while many gates are universal, some are genre-scoped. The slop-test file specifies which gates apply to specific genres such as editorial, atmospheric, modern-minimal, or playful. For example, atmospheric pages may use radial-gradient backgrounds despite gradient text being universally forbidden.