Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Expert in collecting, analyzing, and synthesizing user feedback from multiple channels to extract actionable product insights. Specializes in transforming qualitative feedback into quantitative pri...
.claude/skills/dev-dennis-040-product-feedback-synthesizer/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 26% | 0% |
| case-19 | ✗→✓ | ▲ Improved | -15% | 0% |
| case-20 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-03 | ✓→✗ | ▼ Worse | 35% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | 56% | 0% |
name: Feedback Synthesizer description: Expert in collecting, analyzing, and synthesizing user feedback from multiple channels to extract actionable product insights. Transforms qualitative feedback into quantitative priorities and strategic recommendations. color: blue tools: WebFetch, WebSearch, Read, Write, Edit
Expert in collecting, analyzing, and synthesizing user feedback from multiple channels to extract actionable product insights. Specializes in transforming qualitative feedback into quantitative priorities and strategic recommendations for data-driven product decisions.
Use this agent when you need:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 17,651 | 21,640 | +23% | 1 | 1 | 0% | 4,569 | 7,119 | +56% | 0 | 0 | — |
case-02 | pass→pass | 12,325 | 14,558 | +18% | 1 | 1 | 0% | 2,386 | 3,860 | +62% | 0 | 0 | — |
case-03 | pass→fail | 20,426 | 22,851 | +12% | 1 | 1 | 0% | 4,268 | 5,746 | +35% | 0 | 0 | — |
case-04 | pass→pass | 9,943 | 5,541 | -44% | 1 | 1 | 0% | 1,736 | 2,447 | +41% | 0 | 0 | — |
case-05 | pass→pass | 6,493 | 6,864 | +6% | 1 | 1 | 0% | 1,246 | 2,218 | +78% | 0 | 0 | — |
case-06 | pass→pass | 8,805 | 6,928 | -21% | 1 | 1 | 0% | 1,881 | 2,307 | +23% | 0 | 0 | — |
case-07 | fail→pass | 11,183 | 8,460 | -24% | 1 | 1 | 0% | 2,218 | 2,794 | +26% | 0 | 0 | — |
case-08 | pass→pass | 10,679 | 11,744 | +10% | 1 | 1 | 0% | 2,079 | 3,423 | +65% | 0 | 0 | — |
case-09 | fail→fail | 11,679 | 10,581 | -9% | 1 | 1 | 0% | 2,421 | 3,377 | +39% | 0 | 0 | — |
case-10 | pass→pass | 9,993 | 2,146 | -79% | 1 | 1 | 0% | 1,835 | 1,628 | -11% | 0 | 0 | — |
case-11 | pass→pass | 8,521 | 10,482 | +23% | 1 | 1 | 0% | 1,688 | 3,224 | +91% | 0 | 0 | — |
case-12 | pass→pass | 18,642 | 8,933 | -52% | 1 | 1 | 0% | 3,683 | 2,987 | -19% | 0 | 0 | — |
case-13 | pass→pass | 9,592 | 3,787 | -61% | 1 | 1 | 0% | 1,813 | 1,996 | +10% | 0 | 0 | — |
case-14 | pass→pass | 8,464 | 6,472 | -24% | 1 | 1 | 0% | 1,715 | 2,544 | +48% | 0 | 0 | — |
case-15 | fail→fail | 10,060 | 12,447 | +24% | 1 | 1 | 0% | 2,007 | 3,595 | +79% | 0 | 0 | — |
case-16 | pass→pass | 18,146 | 15,858 | -13% | 1 | 1 | 0% | 2,514 | 3,437 | +37% | 0 | 0 | — |
case-17 | pass→pass | 11,538 | 12,175 | +6% | 1 | 1 | 0% | 2,174 | 3,453 | +59% | 0 | 0 | — |
case-18 | pass→pass | 13,162 | 14,646 | +11% | 1 | 1 | 0% | 2,514 | 3,788 | +51% | 0 | 0 | — |
case-19 | fail→pass | 9,834 | 1,602 | -84% | 1 | 1 | 0% | 1,777 | 1,517 | -15% | 0 | 0 | — |
case-20 | fail→pass | 7,479 | 1,409 | -81% | 1 | 1 | 0% | 1,375 | 1,479 | +8% | 0 | 0 | — |
case-21 | pass→pass | 14,949 | 16,013 | +7% | 1 | 1 | 0% | 2,661 | 4,061 | +53% | 0 | 0 | — |
case-22 | fail→fail | 14,677 | 12,762 | -13% | 1 | 1 | 0% | 2,866 | 3,761 | +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 +9 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is 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.