Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create refined user personas from research data — 3 personas with JTBD, pains, gains, and unexpected insights. Use when building personas from survey data, creating user profiles from research, or segmenting users for product decisions.
.claude/skills/phuryn-user-personas/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 19% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 63% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 27% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 51% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 39% | 0% |
Create detailed, actionable user personas from research data that capture the true diversity of your user base. This skill generates research-backed personas with jobs-to-be-done, pain points, desired outcomes, and unexpected behavioral insights to guide product decisions.
You are an experienced product researcher specializing in persona development and user research synthesis.
Your task is to create 3 refined user personas for $ARGUMENTS.
If the user provides CSV, Excel, survey responses, interview transcripts, or other research data files, read and analyze them directly using available tools. Extract key patterns, demographics, motivations, and behaviors.
For each of the 3 personas, provide:
Persona Name & Demographics
Primary Job-to-be-Done
Top 3 Pain Points
Top 3 Desired Gains
One Unexpected Insight
Product Fit Assessment
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-07 | fail→pass | 19,849 | 20,832 | +5% | 1 | 1 | 0% | 3,330 | 3,970 | +19% | 0 | 0 | — |
case-13 | fail→fail | 21,518 | 22,803 | +6% | 1 | 1 | 0% | 2,868 | 4,366 | +52% | 0 | 0 | — |
case-01 | fail→fail | 19,065 | 29,331 | +54% | 1 | 1 | 0% | 3,297 | 4,282 | +30% | 0 | 0 | — |
case-02 | fail→pass | 14,547 | 27,794 | +91% | 1 | 1 | 0% | 2,471 | 4,020 | +63% | 0 | 0 | — |
case-03 | fail→pass | 17,583 | 19,494 | +11% | 1 | 1 | 0% | 3,014 | 3,817 | +27% | 0 | 0 | — |
case-04 | fail→pass | 12,367 | 15,664 | +27% | 1 | 1 | 0% | 2,066 | 3,111 | +51% | 0 | 0 | — |
case-05 | fail→fail | 15,861 | 22,863 | +44% | 1 | 1 | 0% | 2,686 | 4,309 | +60% | 0 | 0 | — |
case-06 | pass→fail | 19,719 | 21,619 | +10% | 1 | 1 | 0% | 2,482 | 4,168 | +68% | 0 | 0 | — |
case-08 | pass→pass | 16,531 | 23,190 | +40% | 1 | 1 | 0% | 2,611 | 4,445 | +70% | 0 | 0 | — |
case-09 | fail→pass | 17,595 | 22,749 | +29% | 1 | 1 | 0% | 3,100 | 4,299 | +39% | 0 | 0 | — |
case-10 | pass→pass | 18,165 | 27,356 | +51% | 1 | 1 | 0% | 2,699 | 4,065 | +51% | 0 | 0 | — |
case-11 | fail→fail | 14,000 | 26,269 | +88% | 1 | 1 | 0% | 2,336 | 4,047 | +73% | 0 | 0 | — |
case-12 | fail→pass | 18,017 | 21,591 | +20% | 1 | 1 | 0% | 3,302 | 4,079 | +24% | 0 | 0 | — |
case-14 | fail→fail | 16,803 | 22,520 | +34% | 1 | 1 | 0% | 2,526 | 4,326 | +71% | 0 | 0 | — |
case-15 | fail→pass | 17,384 | 19,562 | +13% | 1 | 1 | 0% | 2,754 | 3,681 | +34% | 0 | 0 | — |
case-16 | fail→fail | 14,696 | 24,709 | +68% | 1 | 1 | 0% | 2,562 | 4,599 | +80% | 0 | 0 | — |
case-17 | fail→fail | 17,126 | 24,091 | +41% | 1 | 1 | 0% | 2,911 | 4,509 | +55% | 0 | 0 | — |
case-18 | fail→fail | 16,874 | 30,102 | +78% | 1 | 1 | 0% | 2,417 | 4,357 | +80% | 0 | 0 | — |
case-19 | fail→pass | 19,834 | 18,358 | -7% | 1 | 1 | 0% | 2,495 | 3,202 | +28% | 0 | 0 | — |
case-20 | pass→pass | 16,337 | 12,655 | -23% | 1 | 1 | 0% | 2,253 | 2,809 | +25% | 0 | 0 | — |
case-21 | pass→fail | 18,310 | 22,061 | +20% | 1 | 1 | 0% | 3,389 | 4,608 | +36% | 0 | 0 | — |
case-22 | pass→fail | 14,931 | 24,882 | +67% | 1 | 1 | 0% | 2,894 | 4,743 | +64% | 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 +23 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.
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.