---
path_id: landing-page-visual-audit
path_version: 1
status: partial
generated_from: SkillsTested structured records
---

# Landing page audit, anti-slop pass, and OG brief

> Evidence status: PARTIAL EVIDENCE. This pack does not extend the published evidence boundary.

## Outcome

A prioritized design audit, confirmed anti-pattern report, proposed correction list, and an implementation-ready OG image brief.

## Required inputs

- Public URL, local page, or repository
- Page purpose and intended audience
- Declared design genre or design-system constraints

## Expected outputs

- Severity-ranked visual audit
- Scanner report with human-confirmed findings
- Correction checklist
- Model-routed 1200×630 OG image brief

## Pinned installs

### hallmark

- Source: https://github.com/Nutlope/hallmark
- Commit tested: `aeb42fb`
- Verification: `function_verified`
- Install: `npx skills add nutlope/hallmark -a claude-code -y`
- Not verified: A generated design, repeated-run consistency, or behavior outside Claude Code.

### kill-ai-slop

- Source: https://github.com/yetone/kill-ai-slop
- Commit tested: `88cd544`
- Verification: `function_verified`
- Install: `npx skills add yetone/kill-ai-slop -a claude-code -y`
- Not verified: Comparative quality improvement across repeated real projects.

### visual-skills (video + image)

- Source: https://github.com/smixs/visual-skills
- Commit tested: `75d8ce8`
- Verification: `function_verified`
- Install: `npx skills add smixs/visual-skills -a claude-code -y  (works despite README only documenting manual cp)`
- Not verified: A generated image/video artifact or repeated output-quality comparison.

## Execution

### 1. Audit with judgment

- Purpose: Find hierarchy, genre, typography, and macrostructure problems that mechanical rules cannot judge.
- Input: Rendered page plus its audience, purpose, and declared design genre
- Instruction: Use hallmark's audit verb and record each finding as what, where, severity, and proposed fix without editing the page yet.
- Output: hallmark-audit.md
- Evidence level: `function_verified`
- Human confirmation: required

### 2. Scan and confirm anti-patterns

- Purpose: Catch mechanical regressions quickly, then separate true problems from design-system exemptions and scanner self-matches.
- Input: Page source and declared design-system exemptions
- Instruction: Run the bundled scanner, retain raw hits, and complete its required human-confirm step before proposing fixes.
- Output: anti-slop-report.md
- Evidence level: `function_verified`
- Human confirmation: required

### 3. Brief the social card

- Purpose: Choose an image model based on the content and convert the approved page direction into a constrained OG prompt.
- Input: Approved brand direction, exact copy, palette, and 1200×630 use case
- Instruction: Follow visual-skills' mandatory model-routing reference before producing the five-slot image brief with exact-text constraints.
- Output: og-image-brief.md
- Evidence level: `function_verified`
- Human confirmation: required

### 4. Verify the corrected page and OG asset

- Purpose: Check that fixes did not introduce regressions and that the final image meets its dimensions and copy contract.
- Input: Corrected page and generated OG image
- Instruction: Re-run mechanical checks, capture responsive screenshots, run basic accessibility checks, and inspect OG dimensions, text, and safe margins.
- Output: before-after report and final-og.png
- Evidence level: `unverified`
- Human confirmation: required

## Evidence

### Three design skills applied to the same live homepage

- Date: 2026-07-25
- What it proves: hallmark produced a severity-ranked audit, kill-ai-slop ran its scanner and human-confirm pass, and visual-skills produced a model-routed 1200 by 630 social-card brief for the same page.
- Not verified: Applying the proposed fixes, generating a final OG image, repeated-run consistency, or measuring quality gain against a no-skill baseline.
- Source: https://skillstested.com/compare/hallmark-vs-kill-ai-slop-vs-visual-skills/

## Known limits

- The published run audited the page but did not apply the fixes.
- The social-card deliverable is a prompt brief, not a generated image.
- hallmark's genre severity is ambiguous when a page has no design stamp.
- kill-ai-slop's raw scanner hits require human confirmation.

## Next verification

Freeze one deliberately flawed landing page, record a no-skill baseline, apply all three stages, publish before and after screenshots plus a real OG image, and repeat the run three times.