Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Specify micro-interactions with trigger, rules, feedback, and loop/mode definitions.
.claude/skills/owl-listener-micro-interaction-spec/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✓→✗ | ▼ Worse | 16% | 0% |
| case-21 | ✓→✗ | ▼ Worse | 2% | 0% |
| case-22 | ✓→✗ | ▼ Worse | 86% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | -10% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 18% | 0% |
You are an expert in designing micro-interactions that make interfaces feel alive and intuitive.
You specify micro-interactions using a structured framework covering trigger, rules, feedback, and loops.
What initiates the interaction: user action (click, hover, swipe), system event (notification, completion), or conditional (time-based, threshold).
What happens once triggered: the logic and sequence of the interaction, conditions and branching.
How the user perceives the result: visual change (color, size, position), motion (animation, transition), audio (click, chime), haptic (vibration patterns).
Does the interaction repeat? Does it change over time? First-time vs repeat behavior, progressive disclosure.
For each micro-interaction: name, trigger, rules (sequence), feedback (visual/audio/haptic), duration/easing, loop behavior, accessibility considerations.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 21,364 | 16,486 | -23% | 1 | 1 | 0% | 3,691 | 3,229 | -13% | 0 | 0 | — |
case-02 | pass→pass | 19,435 | 15,333 | -21% | 1 | 1 | 0% | 3,428 | 3,086 | -10% | 0 | 0 | — |
case-03 | pass→fail | 15,298 | 16,132 | +5% | 1 | 1 | 0% | 2,691 | 3,116 | +16% | 0 | 0 | — |
case-04 | pass→pass | 14,343 | 14,936 | +4% | 1 | 1 | 0% | 2,496 | 2,940 | +18% | 0 | 0 | — |
case-05 | fail→fail | 19,530 | 13,538 | -31% | 1 | 1 | 0% | 3,549 | 2,589 | -27% | 0 | 0 | — |
case-15 | fail→fail | 15,655 | 14,065 | -10% | 1 | 1 | 0% | 3,397 | 3,151 | -7% | 0 | 0 | — |
case-06 | fail→fail | 16,826 | 12,925 | -23% | 1 | 1 | 0% | 3,023 | 2,639 | -13% | 0 | 0 | — |
case-07 | pass→pass | 17,450 | 12,445 | -29% | 1 | 1 | 0% | 2,834 | 2,523 | -11% | 0 | 0 | — |
case-08 | fail→fail | 15,496 | 15,663 | +1% | 1 | 1 | 0% | 2,751 | 3,128 | +14% | 0 | 0 | — |
case-09 | fail→fail | 21,390 | 15,284 | -29% | 1 | 1 | 0% | 3,851 | 2,855 | -26% | 0 | 0 | — |
case-10 | fail→fail | 25,375 | 15,972 | -37% | 1 | 1 | 0% | 4,830 | 3,040 | -37% | 0 | 0 | — |
case-11 | fail→fail | 14,394 | 15,921 | +11% | 1 | 1 | 0% | 2,408 | 2,760 | +15% | 0 | 0 | — |
case-12 | fail→fail | 10,117 | 11,154 | +10% | 1 | 1 | 0% | 1,900 | 2,246 | +18% | 0 | 0 | — |
case-13 | fail→fail | 9,413 | 11,243 | +19% | 1 | 1 | 0% | 1,507 | 2,207 | +46% | 0 | 0 | — |
case-14 | fail→fail | 14,501 | 16,193 | +12% | 1 | 1 | 0% | 2,537 | 3,340 | +32% | 0 | 0 | — |
case-16 | pass→pass | 17,912 | 11,070 | -38% | 1 | 1 | 0% | 3,906 | 2,608 | -33% | 0 | 0 | — |
case-17 | fail→fail | 12,573 | 11,584 | -8% | 1 | 1 | 0% | 2,351 | 2,447 | +4% | 0 | 0 | — |
case-18 | fail→fail | 16,206 | 12,729 | -21% | 1 | 1 | 0% | 2,838 | 2,763 | -3% | 0 | 0 | — |
case-19 | pass→pass | 18,301 | 17,701 | -3% | 1 | 1 | 0% | 3,589 | 3,254 | -9% | 0 | 0 | — |
case-20 | pass→pass | 10,375 | 6,632 | -36% | 1 | 1 | 0% | 1,930 | 1,479 | -23% | 0 | 0 | — |
case-21 | pass→fail | 11,106 | 10,085 | -9% | 1 | 1 | 0% | 2,044 | 2,094 | +2% | 0 | 0 | — |
case-22 | pass→fail | 4,867 | 7,738 | +59% | 1 | 1 | 0% | 1,151 | 2,140 | +86% | 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. 3 cases got worse with the skill loaded, and they are included in that figure.
Other measured skills in the registry, with their headline benchmark lift.