Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Google Ads deep analysis covering Search, Performance Max, Display, YouTube, and Demand Gen campaigns. Evaluates 74 checks across conversion tracking, wasted spend, account structure, keywords, ads, and settings. Triggers on: "Google Ads", "Google PPC", "search ads", "PMax", "Performance Max", "Google campaign", "Google audit"
.claude/skills/miosa-osa-ads-google/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 109% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 73% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 39% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 55% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 31% | 0% |
> Comprehensive 74-check audit of Google Ads accounts across all campaign types.
bash/ads google /ads google --focus pmax
Negative Keyword Rules (critical):
[keyword] for specific irrelevant queries"keyword" for irrelevant intent patternsIf Performance Max campaigns exist, additionally evaluate:
If AI Max for Search is available/active:
| Metric | Pass | Warning | Fail | |--------|------|---------|------| | Quality Score (avg) | >=7 | 5-6 | <5 | | CTR (Search) | >=6.66% | 3-6.66% | <3% | | CVR (Search) | >=7.52% | 3-7.52% | <3% | | CPC (Search) | <=$5.26 | $5.26-8.00 | >$8.00 | | Wasted Spend | <10% | 10-20% | >20% | | Ad Strength | Good+ | Average | Poor | | Invalid Clicks | <5% | 5-10% | >10% |
GOOGLE-ADS-REPORT.md — Full 74-check findings with pass/warning/fail| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 19,177 | 56,990 | +197% | 1 | 1 | 0% | 3,434 | 9,361 | +173% | 0 | 0 | — |
case-02 | fail→fail | 25,456 | 42,618 | +67% | 1 | 1 | 0% | 4,134 | 8,957 | +117% | 0 | 0 | — |
case-03 | fail→pass | 26,795 | 46,930 | +75% | 1 | 1 | 0% | 4,478 | 9,356 | +109% | 0 | 0 | — |
case-04 | fail→pass | 12,495 | 14,718 | +18% | 1 | 1 | 0% | 2,030 | 3,503 | +73% | 0 | 0 | — |
case-05 | fail→pass | 13,265 | 10,912 | -18% | 1 | 1 | 0% | 2,115 | 2,931 | +39% | 0 | 0 | — |
case-06 | pass→pass | 9,723 | 10,959 | +13% | 1 | 1 | 0% | 1,569 | 2,830 | +80% | 0 | 0 | — |
case-07 | fail→pass | 11,187 | 10,113 | -10% | 1 | 1 | 0% | 1,889 | 2,922 | +55% | 0 | 0 | — |
case-08 | fail→pass | 10,215 | 6,691 | -34% | 1 | 1 | 0% | 1,661 | 2,176 | +31% | 0 | 0 | — |
case-09 | fail→pass | 11,134 | 7,939 | -29% | 1 | 1 | 0% | 1,936 | 2,310 | +19% | 0 | 0 | — |
case-10 | fail→pass | 15,505 | 11,003 | -29% | 1 | 1 | 0% | 2,426 | 2,888 | +19% | 0 | 0 | — |
case-11 | pass→pass | 9,893 | 5,888 | -40% | 1 | 1 | 0% | 1,773 | 1,996 | +13% | 0 | 0 | — |
case-12 | fail→pass | 8,224 | 6,827 | -17% | 1 | 1 | 0% | 1,159 | 2,266 | +96% | 0 | 0 | — |
case-13 | fail→pass | 11,793 | 10,126 | -14% | 1 | 1 | 0% | 1,900 | 2,725 | +43% | 0 | 0 | — |
case-14 | pass→pass | 6,501 | 7,200 | +11% | 1 | 1 | 0% | 882 | 2,291 | +160% | 0 | 0 | — |
case-15 | pass→pass | 15,915 | 11,748 | -26% | 1 | 1 | 0% | 2,301 | 2,693 | +17% | 0 | 0 | — |
case-16 | pass→pass | 6,311 | 1,854 | -71% | 1 | 1 | 0% | 950 | 1,403 | +48% | 0 | 0 | — |
case-17 | pass→pass | 11,106 | 10,872 | -2% | 1 | 1 | 0% | 1,558 | 3,030 | +94% | 0 | 0 | — |
case-18 | fail→pass | 17,453 | 11,726 | -33% | 1 | 1 | 0% | 2,220 | 3,002 | +35% | 0 | 0 | — |
case-19 | fail→pass | 20,491 | 1,556 | -92% | 1 | 1 | 0% | 2,896 | 1,388 | -52% | 0 | 0 | — |
case-20 | pass→pass | 19,846 | 22,871 | +15% | 1 | 1 | 0% | 2,787 | 4,516 | +62% | 0 | 0 | — |
case-21 | pass→pass | 4,632 | 18,466 | +299% | 1 | 1 | 0% | 669 | 4,212 | +530% | 0 | 0 | — |
case-22 | pass→pass | 22,508 | 22,903 | +2% | 1 | 1 | 0% | 3,814 | 5,000 | +31% | 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 +50 percentage points is the difference between those two pass rates over the 22 comparable cases.
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.