Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Identify emerging trends and viral topics relevant to your niche before they peak.
.claude/skills/holaboss-ai-trend-spotter/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -15% | 0% |
| case-01 | ✗→✓ | ▲ Improved | 3% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 21% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 14% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 14% | 0% |
Think like an editor with a finger on the pulse: your value is catching a rising topic while it's still climbing, not reporting one that already peaked. The goal is timing and actionable angles, not a list of yesterday's news.
Use Trend Spotter to find what's gaining momentum in a specific niche: emerging topics, accelerating hashtags, formats catching on, conversations starting to spike. If the task is to deeply research an already-established topic, use Web Researcher; if it's to turn signals into a slate of content concepts, hand off to Idea Generator.
Distinguish a trend from noise and from a fad:
Be honest when something is a flash in the pan or already over — recommending a dead trend wastes the user's effort.
Return a ranked shortlist of trends, strongest opportunity first. For each:
Lead with timing-sensitive opportunities. Note clearly when a candidate is speculative or already cresting.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | fail→pass | 26,627 | 19,948 | -25% | 1 | 1 | 0% | 3,734 | 3,170 | -15% | 0 | 0 | — |
case-01 | fail→pass | 21,190 | 20,824 | -2% | 1 | 1 | 0% | 3,050 | 3,138 | +3% | 0 | 0 | — |
case-03 | fail→fail | 18,492 | 16,434 | -11% | 1 | 1 | 0% | 2,794 | 2,813 | +1% | 0 | 0 | — |
case-04 | fail→fail | 27,779 | 22,199 | -20% | 1 | 1 | 0% | 4,570 | 3,750 | -18% | 0 | 0 | — |
case-05 | fail→fail | 41,319 | 34,103 | -17% | 1 | 1 | 0% | 5,789 | 5,785 | -0% | 0 | 0 | — |
case-06 | pass→fail | 52,314 | 28,765 | -45% | 1 | 1 | 0% | 8,213 | 5,279 | -36% | 0 | 0 | — |
case-07 | pass→pass | 8,924 | 9,198 | +3% | 1 | 1 | 0% | 1,472 | 1,809 | +23% | 0 | 0 | — |
case-08 | fail→pass | 17,679 | 17,626 | -0% | 1 | 1 | 0% | 2,579 | 3,126 | +21% | 0 | 0 | — |
case-09 | fail→pass | 13,814 | 10,722 | -22% | 1 | 1 | 0% | 1,817 | 2,063 | +14% | 0 | 0 | — |
case-10 | fail→pass | 21,563 | 22,103 | +3% | 1 | 1 | 0% | 3,394 | 3,876 | +14% | 0 | 0 | — |
case-11 | fail→fail | 31,108 | 26,152 | -16% | 1 | 1 | 0% | 4,223 | 4,312 | +2% | 0 | 0 | — |
case-12 | fail→pass | 18,415 | 16,697 | -9% | 1 | 1 | 0% | 2,611 | 2,682 | +3% | 0 | 0 | — |
case-13 | pass→pass | 12,883 | 12,217 | -5% | 1 | 1 | 0% | 1,913 | 2,214 | +16% | 0 | 0 | — |
case-14 | fail→fail | 17,864 | 15,024 | -16% | 1 | 1 | 0% | 2,353 | 2,429 | +3% | 0 | 0 | — |
case-15 | fail→fail | 19,199 | 18,895 | -2% | 1 | 1 | 0% | 2,743 | 2,864 | +4% | 0 | 0 | — |
case-16 | pass→pass | 10,574 | 12,675 | +20% | 1 | 1 | 0% | 1,761 | 1,913 | +9% | 0 | 0 | — |
case-17 | fail→pass | 24,674 | 21,044 | -15% | 1 | 1 | 0% | 3,882 | 3,216 | -17% | 0 | 0 | — |
case-18 | pass→pass | 12,181 | 11,011 | -10% | 1 | 1 | 0% | 1,799 | 2,105 | +17% | 0 | 0 | — |
case-19 | fail→pass | 24,810 | 19,609 | -21% | 1 | 1 | 0% | 2,991 | 3,073 | +3% | 0 | 0 | — |
case-20 | fail→pass | 20,759 | 16,633 | -20% | 1 | 1 | 0% | 2,691 | 2,755 | +2% | 0 | 0 | — |
case-21 | pass→pass | 19,577 | 13,608 | -30% | 1 | 1 | 0% | 2,635 | 2,223 | -16% | 0 | 0 | — |
case-22 | fail→pass | 25,399 | 24,647 | -3% | 1 | 1 | 0% | 3,676 | 3,807 | +4% | 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 +41 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.