{
  "path": {
    "schema_version": "1.0.0",
    "path_id": "landing-page-visual-audit",
    "path_version": 1,
    "status": "partial",
    "commercial_ready": true,
    "title": "Landing page audit, implementation patch, and final OG",
    "short_title": "Landing page → visual QA",
    "summary": "Use hallmark for judgment, kill-ai-slop for mechanical recall, and visual-skills for social-card direction, then deliver an inspectable page patch and final OG image.",
    "outcome": "A prioritized design audit, raw and human-reviewed scanner report, applied page patch, and final 1200×630 OG PNG.",
    "match_terms": [
      {
        "term": "landing page",
        "weight": 8
      },
      {
        "term": "website audit",
        "weight": 8
      },
      {
        "term": "design audit",
        "weight": 8
      },
      {
        "term": "site review",
        "weight": 7
      },
      {
        "term": "ux audit",
        "weight": 8
      },
      {
        "term": "homepage",
        "weight": 5
      },
      {
        "term": "web design",
        "weight": 6
      },
      {
        "term": "social card",
        "weight": 6
      },
      {
        "term": "design critique",
        "weight": 7
      },
      {
        "term": "ai-looking",
        "weight": 7
      },
      {
        "term": "ai slop",
        "weight": 7
      },
      {
        "term": "og image",
        "weight": 6
      },
      {
        "term": "website",
        "weight": 3
      },
      {
        "term": "ui",
        "weight": 2
      },
      {
        "term": "落地页",
        "weight": 8
      },
      {
        "term": "官网",
        "weight": 5
      },
      {
        "term": "设计审计",
        "weight": 8
      },
      {
        "term": "页面审查",
        "weight": 8
      },
      {
        "term": "ui审计",
        "weight": 8
      },
      {
        "term": "网页设计",
        "weight": 6
      },
      {
        "term": "首页",
        "weight": 5
      },
      {
        "term": "社交分享图",
        "weight": 6
      },
      {
        "term": "去ai味",
        "weight": 7
      },
      {
        "term": "去 ai 味",
        "weight": 7
      },
      {
        "term": "og图",
        "weight": 6
      },
      {
        "term": "网页",
        "weight": 3
      }
    ],
    "supported_agents": [
      "claude-code"
    ],
    "starting_points": [
      "url",
      "repo",
      "files"
    ],
    "priority_fit": {
      "best-quality": 5,
      "fastest": 3,
      "leanest": 4,
      "most-private": 4
    },
    "required_inputs": [
      "Public URL, local page, or repository",
      "Page purpose and intended audience",
      "Declared design genre or design-system constraints"
    ],
    "outputs": [
      "Severity-ranked visual audit",
      "Scanner report with human-confirmed findings",
      "Applied static-page patch and correction checklist",
      "Final 1200×630 OG PNG"
    ],
    "required_output_types": [
      "text/markdown",
      "image/png"
    ],
    "nodes": [
      {
        "node_id": "judgment-audit",
        "node_type": "skill",
        "stage": "Audit with judgment",
        "tested_ref": "hallmark",
        "purpose": "Find hierarchy, genre, typography, and macrostructure problems that mechanical rules cannot judge.",
        "input": "Rendered page plus its audience, purpose, and declared design genre",
        "output": "hallmark-audit.md",
        "verification": "function_verified",
        "instruction": "Use hallmark's audit verb and record each finding as what, where, severity, and proposed fix without editing the page yet.",
        "human_confirmation": true
      },
      {
        "node_id": "mechanical-scan",
        "node_type": "skill",
        "stage": "Scan and confirm anti-patterns",
        "tested_ref": "kill-ai-slop",
        "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",
        "output": "anti-slop-report.md",
        "verification": "function_verified",
        "instruction": "Run the bundled scanner, retain raw hits, and complete its required human-confirm step before proposing fixes.",
        "human_confirmation": true
      },
      {
        "node_id": "og-brief",
        "node_type": "skill",
        "stage": "Create the social card direction",
        "tested_ref": "visual-skills",
        "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",
        "output": "og-image-direction.md",
        "verification": "function_verified",
        "instruction": "Follow visual-skills' mandatory model-routing reference, freeze the exact copy and safe margins, then turn the approved direction into a final vector-first 1200×630 asset.",
        "human_confirmation": true
      },
      {
        "node_id": "final-verification",
        "node_type": "verify",
        "stage": "Verify the corrected page and OG asset",
        "tested_ref": null,
        "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",
        "output": "before-after report and final-og.png",
        "verification": "workflow_verified",
        "instruction": "Re-run mechanical and structural checks, confirm the analytics and privacy scripts are byte-preserved, and inspect OG dimensions, text, and safe margins.",
        "human_confirmation": true
      }
    ],
    "evidence_refs": [
      "design-comparison-2026-07-25",
      "landing-reproduction-2026-08-02"
    ],
    "known_limits": [
      "The nine-run reproduction used one frozen static homepage, not nine different customer stacks.",
      "The service applies a patch only to supplied static source; it does not deploy to production.",
      "hallmark's genre severity is ambiguous when a page has no design stamp.",
      "kill-ai-slop's raw scanner hits require human confirmation, and scanner counts did not distinguish the three experiment arms."
    ],
    "next_test": "Fulfil the first non-owner order within the promised three-business-day window and record rework, refund, and buyer acceptance without publishing private inputs."
  },
  "evidence": [
    {
      "schema_version": "1.0.0",
      "evidence_id": "design-comparison-2026-07-25",
      "evidence_type": "comparison-run",
      "date": "2026-07-25",
      "title": "Three design skills applied to the same live homepage",
      "scope": "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_ref": "hallmark-vs-kill-ai-slop-vs-visual-skills",
      "source_url": "/compare/hallmark-vs-kill-ai-slop-vs-visual-skills/",
      "environment": {
        "os": "Linux 6.8, arm64, 2 cores",
        "node": "v22.22.3"
      },
      "metrics": [
        "hallmark: 0 critical, 1 major, 3 minor findings before genre confirmation",
        "kill-ai-slop: 30 raw hits across 3 groups, about 1 surviving human confirmation",
        "visual-skills: one structured GPT Image social-card brief"
      ],
      "artifacts": [
        {
          "label": "Full design comparison and raw findings",
          "type": "page",
          "path": "/compare/hallmark-vs-kill-ai-slop-vs-visual-skills/",
          "alt": "Open the published three-skill design comparison"
        }
      ]
    },
    {
      "schema_version": "1.0.0",
      "evidence_id": "landing-reproduction-2026-08-02",
      "evidence_type": "comparison-run",
      "date": "2026-08-02",
      "title": "Nine independent landing-page delivery runs with final OG assets",
      "scope": "The same production homepage was run three times with no skill context, Hallmark alone, and the published three-skill combination. All nine runs produced a revised full HTML page, substantive audit, and renderable 1200 by 630 OG; all nine preserved the analytics and goal-privacy scripts byte for byte and passed the same structural delivery checks.",
      "not_verified": "That the skill combination is higher quality than the control, that arbitrary customer frameworks need no rework, or that a paid order can always be fulfilled inside three business days. Scanner medians were identical across arms.",
      "source_ref": "hallmark-vs-kill-ai-slop-vs-visual-skills",
      "environment": {
        "model": "gpt-5.6-terra",
        "reasoning_effort": "medium",
        "input_sha256": "6d80aa1364afee07a955decc205014b505e9a2c2167d2badc79c1872c060afb4",
        "runs": 9
      },
      "metrics": [
        "Valid complete deliveries: 9 of 9",
        "Median duration: control 126.30s, single skill 229.87s, combination 168.14s",
        "Median scanner result in every arm: 4 groups and 26 raw hits",
        "Selected final asset: combination run 2, chosen for clearest product-path communication rather than a claimed arm-level quality win"
      ],
      "artifacts": [
        {
          "label": "Selected final OG image",
          "type": "image",
          "path": "/assets/evidence/landing-reproduction-og.png",
          "alt": "SkillsTested social card showing outcome, path, and evidence as a three-step flow"
        },
        {
          "label": "Nine-run mechanical summary",
          "type": "record",
          "path": "/assets/evidence/landing-reproduction-summary.json",
          "alt": "Machine-readable results for all three experiment arms"
        }
      ]
    }
  ],
  "tested_skills": {
    "hallmark": {
      "skill": "hallmark",
      "repo": "Nutlope/hallmark",
      "stars": 17177,
      "stars_date": "2026-07-25",
      "commit_hash": "aeb42fb",
      "version_tested": "1.1.0",
      "install_cmd": "npx skills add nutlope/hallmark -a claude-code -y",
      "env": {
        "os": "Linux 6.18.5 (cloud sandbox)",
        "node": "v22.22.2"
      },
      "clone_time_sec": 1.7,
      "install_time_sec": 2.5,
      "failures": [],
      "verification_level": "function_verified",
      "verified_scope": "Installation plus a same-page audit that produced a severity-ranked punch list in the design comparison.",
      "not_verified": "A generated design, repeated-run consistency, or behavior outside Claude Code.",
      "verification_evidence": "/compare/hallmark-vs-kill-ai-slop-vs-visual-skills/",
      "min_test": {
        "what": "install + structure inspect",
        "result": "SKILL.md v1.1.0 + references/ land in .claude/skills/hallmark; instruction-only, no runtime deps",
        "time_sec": 2.5
      },
      "notes": "Anti-AI-slop design skill by Nutlope (Hassan El Mghari, vercel). Direct competitor-pair page material vs kill-ai-slop: hallmark = design-system generation angle, kill-ai-slop = detect-and-fix angle with scanner script. A tested comparison page of the two is a natural /skills/design/ section.",
      "tested_date": "2026-07-23",
      "skills_installed": 1,
      "stars_note": "15,794 (07-23) -> 17,177 (07-25), re-read for design deep comparison."
    },
    "kill-ai-slop": {
      "skill": "kill-ai-slop",
      "repo": "yetone/kill-ai-slop",
      "stars": 738,
      "stars_date": "2026-07-25",
      "commit_hash": "88cd544",
      "install_cmd": "npx skills add yetone/kill-ai-slop -a claude-code -y",
      "install_alt": "manual: copy skill/ into .claude/skills/",
      "env": {
        "os": "Linux 6.18.5 (cloud sandbox)",
        "node": "v22.22.2",
        "npm": "10.9.7",
        "python": "3.11.15"
      },
      "install_time_sec": 2.8,
      "npx_cli_first_run_sec": 4.7,
      "clone_time_sec": 1.7,
      "failures": [],
      "verification_level": "function_verified",
      "verified_scope": "The bundled scanner executed against a deliberately planted HTML fixture and returned file-level findings.",
      "not_verified": "Comparative quality improvement across repeated real projects.",
      "min_test": {
        "what": "scan.mjs on a deliberately sloppy HTML page",
        "result": "7 slop groups / 8 hits detected (indigo gradient, gradient-clip headline, AI voice, emoji, glassmorphism, corner nesting, invented stats)",
        "time_sec": 0.06
      },
      "structure": "SKILL.md + references/(detection|fixes|taxonomy).md + scripts/scan.mjs(+tests)",
      "notes": "Author runs killaislop.com (Astro field guide, 33 tells, 4 languages). Skill and site share one taxonomy source. Cleanest install of the batch. Install path = universal 'npx skills' CLI (vercel-labs), non-interactive flags -a/-s/-y work.",
      "tested_date": "2026-07-23",
      "skills_installed": 1,
      "stars_note": "683 (07-22) / 708 (07-23) / 738 (07-25), re-read for design deep comparison."
    },
    "visual-skills": {
      "skill": "visual-skills (video + image)",
      "repo": "smixs/visual-skills",
      "stars": 108,
      "stars_note": "103 (07-23) -> 108 (07-25), re-read for design deep comparison.",
      "commit_hash": "75d8ce8",
      "install_cmd": "npx skills add smixs/visual-skills -a claude-code -y  (works despite README only documenting manual cp)",
      "install_alt": "git clone + cp -r video image ~/.claude/skills/ (documented path, 0.005s)",
      "env": {
        "os": "Linux 6.18.5 (cloud sandbox)",
        "node": "v22.22.2"
      },
      "clone_time_sec": 0.9,
      "install_time_sec": 2.5,
      "failures": [],
      "verification_level": "function_verified",
      "verified_scope": "Both install paths worked; the image skill produced a model-routed, structured social-card brief in the design comparison.",
      "not_verified": "A generated image/video artifact or repeated output-quality comparison.",
      "verification_evidence": "/compare/hallmark-vs-kill-ai-slop-vs-visual-skills/",
      "min_test": {
        "what": "both install paths + structure inspect",
        "result": "video/ and image/ skills land; instruction-only prompt-writing skills for Nano Banana/GPT Image/Seedance/Kling/Veo",
        "time_sec": 2.5
      },
      "notes": "README usage examples are in Russian — worth a callout for EN readers. Covers multi-model video/image prompting.",
      "tested_date": "2026-07-23",
      "stars_date": "2026-07-25",
      "skills_installed": 2
    }
  }
}