Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Respond to Google, Yelp, and industry reviews in the owner's voice -- gracious on the 5-stars, masterful on the 1-stars. Handles the angry customer, the unfair review, the fake review, and the one that mentions a legal or health issue. Every response written for the thousand future customers reading it.
.claude/skills/onewave-ai-review-response-writer/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 43% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 16% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 52% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 48% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 106% | 0% |
The response to a bad review is not for the reviewer -- it is for every prospect who reads it for the next five years. Input: the review text (pasted, exported, or fetched from the profile), and on first run a few existing responses or a note on how the owner talks. Output: responses ready to post, plus escalation flags for the ones that need more than words.
GLOW (4-5 stars) -- thank specifically (echo the detail they praised, never generic), reinforce the service mentioned (it is searchable text), invite them back.LEGITIMATE COMPLAINT -- acknowledge the specific failure without excuses, state the fix made, take it offline with a real contact, never argue. No coupons in public (it trains complaint-for-discount).UNFAIR/MISTAKEN -- correct the record factually and briefly ("Our records show we honored the quoted price of...") while staying gracious; readers can tell who is being reasonable.SUSPECTED FAKE (no record of the customer, competitor patterns) -- respond once, neutrally ("We have no record of serving you -- contact us and we'll make it right"), and output the platform's removal-request steps with the policy grounds.ESCALATE -- anything mentioning injury, illness, discrimination, legal threats, or an employee by name in an accusation: draft nothing final; flag for the owner and, where serious, counsel, with a holding-pattern response option.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 16,386 | 13,878 | -15% | 1 | 1 | 0% | 2,569 | 2,867 | +12% | 0 | 0 | — |
case-02 | pass→pass | 8,664 | 8,801 | +2% | 1 | 1 | 0% | 1,443 | 2,051 | +42% | 0 | 0 | — |
case-03 | fail→pass | 7,705 | 7,395 | -4% | 1 | 1 | 0% | 1,303 | 1,865 | +43% | 0 | 0 | — |
case-04 | fail→pass | 12,438 | 9,547 | -23% | 1 | 1 | 0% | 1,769 | 2,051 | +16% | 0 | 0 | — |
case-05 | fail→pass | 8,962 | 8,832 | -1% | 1 | 1 | 0% | 1,356 | 2,058 | +52% | 0 | 0 | — |
case-06 | fail→pass | 7,538 | 7,101 | -6% | 1 | 1 | 0% | 1,262 | 1,870 | +48% | 0 | 0 | — |
case-07 | fail→pass | 8,079 | 12,424 | +54% | 1 | 1 | 0% | 1,259 | 2,596 | +106% | 0 | 0 | — |
case-08 | fail→pass | 25,839 | 12,591 | -51% | 1 | 1 | 0% | 5,090 | 2,454 | -52% | 0 | 0 | — |
case-09 | pass→pass | 10,887 | 10,968 | +1% | 1 | 1 | 0% | 1,559 | 2,429 | +56% | 0 | 0 | — |
case-10 | pass→pass | 9,946 | 5,408 | -46% | 1 | 1 | 0% | 1,571 | 1,582 | +1% | 0 | 0 | — |
case-11 | pass→pass | 9,653 | 10,945 | +13% | 1 | 1 | 0% | 1,512 | 2,515 | +66% | 0 | 0 | — |
case-12 | pass→pass | 7,063 | 7,061 | -0% | 1 | 1 | 0% | 1,196 | 1,893 | +58% | 0 | 0 | — |
case-13 | pass→pass | 12,474 | 8,626 | -31% | 1 | 1 | 0% | 1,911 | 2,110 | +10% | 0 | 0 | — |
case-14 | pass→pass | 10,834 | 9,875 | -9% | 1 | 1 | 0% | 1,662 | 2,151 | +29% | 0 | 0 | — |
case-15 | pass→pass | 12,358 | 8,616 | -30% | 1 | 1 | 0% | 1,902 | 2,045 | +8% | 0 | 0 | — |
case-16 | fail→pass | 10,845 | 7,102 | -35% | 1 | 1 | 0% | 1,751 | 1,830 | +5% | 0 | 0 | — |
case-17 | fail→pass | 3,541 | 9,344 | +164% | 1 | 1 | 0% | 685 | 2,279 | +233% | 0 | 0 | — |
case-18 | pass→pass | 9,839 | 8,395 | -15% | 1 | 1 | 0% | 1,469 | 1,951 | +33% | 0 | 0 | — |
case-19 | pass→pass | 3,834 | 4,515 | +18% | 1 | 1 | 0% | 639 | 1,460 | +128% | 0 | 0 | — |
case-20 | pass→pass | 10,565 | 9,184 | -13% | 1 | 1 | 0% | 1,582 | 2,170 | +37% | 0 | 0 | — |
case-21 | fail→pass | 11,816 | 11,585 | -2% | 1 | 1 | 0% | 1,768 | 2,499 | +41% | 0 | 0 | — |
case-22 | fail→pass | 16,117 | 11,555 | -28% | 1 | 1 | 0% | 2,407 | 2,383 | -1% | 0 | 0 | — |
case-23 | fail→pass | 19,087 | 13,571 | -29% | 1 | 1 | 0% | 2,971 | 2,426 | -18% | 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. 23 cases were attempted. The headline lift of +48 percentage points is the difference between those two pass rates over the 23 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.