Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Audit a website for SEO issues and track improvements over time. Use when analyzing website SEO performance.
.claude/skills/skrun-dev-seo-audit/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | -35% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -24% | 0% |
| case-01 | ✗→✓ | ▲ Improved | -18% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 28% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -19% | 0% |
You are an expert SEO auditor. Analyze the provided website URL and produce a comprehensive SEO audit report.
Return a JSON object with:
seo_report: An object containing the detailed audit findingsscore: A number 0-100 representing overall SEO healthprevious_score: The score from the last audit (0 if first run)trend: "improving", "declining", "stable", or "first_audit"If you receive previous state, use it to:
Include a _state field in your response with the current audit data to persist for the next run:
json{ "_state": { "last_score": 85, "last_audit_date": "2026-03-22", "audit_count": 2 } }
First run: { "score": 72, "previous_score": 0, "trend": "first_audit" } Second run: { "score": 85, "previous_score": 72, "trend": "improving" }
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-08 | fail→pass | 14,982 | 7,403 | -51% | 1 | 1 | 0% | 2,642 | 1,719 | -35% | 0 | 0 | — |
case-07 | fail→pass | 14,745 | 8,133 | -45% | 1 | 1 | 0% | 2,644 | 1,998 | -24% | 0 | 0 | — |
case-01 | fail→pass | 16,751 | 11,927 | -29% | 1 | 1 | 0% | 3,259 | 2,662 | -18% | 0 | 0 | — |
case-02 | fail→pass | 8,298 | 10,077 | +21% | 1 | 1 | 0% | 1,717 | 2,205 | +28% | 0 | 0 | — |
case-03 | fail→pass | 13,507 | 9,328 | -31% | 1 | 1 | 0% | 2,868 | 2,312 | -19% | 0 | 0 | — |
case-04 | pass→fail | 3,387 | 8,499 | +151% | 1 | 1 | 0% | 649 | 2,103 | +224% | 0 | 0 | — |
case-05 | pass→fail | 7,876 | 7,949 | +1% | 1 | 1 | 0% | 1,759 | 1,952 | +11% | 0 | 0 | — |
case-06 | pass→fail | 6,111 | 8,395 | +37% | 1 | 1 | 0% | 1,412 | 2,133 | +51% | 0 | 0 | — |
case-09 | fail→pass | 7,789 | 6,677 | -14% | 1 | 1 | 0% | 1,495 | 1,690 | +13% | 0 | 0 | — |
case-10 | fail→pass | 14,866 | 7,559 | -49% | 1 | 1 | 0% | 2,415 | 1,694 | -30% | 0 | 0 | — |
case-11 | fail→pass | 15,388 | 9,132 | -41% | 1 | 1 | 0% | 3,158 | 2,285 | -28% | 0 | 0 | — |
case-12 | fail→pass | 16,916 | 10,090 | -40% | 1 | 1 | 0% | 2,580 | 2,453 | -5% | 0 | 0 | — |
case-13 | pass→pass | 10,942 | 8,544 | -22% | 1 | 1 | 0% | 2,161 | 2,142 | -1% | 0 | 0 | — |
case-14 | fail→pass | 13,723 | 8,838 | -36% | 1 | 1 | 0% | 2,435 | 2,060 | -15% | 0 | 0 | — |
case-15 | fail→pass | 11,507 | 9,325 | -19% | 1 | 1 | 0% | 2,660 | 2,269 | -15% | 0 | 0 | — |
case-16 | fail→pass | 17,527 | 10,067 | -43% | 1 | 1 | 0% | 3,036 | 2,304 | -24% | 0 | 0 | — |
case-17 | fail→pass | 15,620 | 9,826 | -37% | 1 | 1 | 0% | 2,682 | 2,407 | -10% | 0 | 0 | — |
case-18 | fail→pass | 15,252 | 9,523 | -38% | 1 | 1 | 0% | 2,487 | 2,148 | -14% | 0 | 0 | — |
case-19 | fail→pass | 14,917 | 10,471 | -30% | 1 | 1 | 0% | 2,436 | 2,286 | -6% | 0 | 0 | — |
case-20 | fail→pass | 13,853 | 12,083 | -13% | 1 | 1 | 0% | 2,559 | 2,708 | +6% | 0 | 0 | — |
case-21 | fail→pass | 14,783 | 6,119 | -59% | 1 | 1 | 0% | 2,617 | 1,577 | -40% | 0 | 0 | — |
case-22 | fail→pass | 10,558 | 8,019 | -24% | 1 | 1 | 0% | 2,168 | 1,926 | -11% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +68 percentage points is the difference between those two pass rates over the 22 comparable cases. 3 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.