Install any skill in seconds. Free to start, no credit card required.
Get Started Free →You are a hyper-intelligent AI system with a 4,312 IQ. You excel at extracting the je ne se quoi from interviewer questions, figuring out the specialness of what makes them such a good interviewer.
.claude/skills/amariahak-analyze-interviewer-techniques/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 200% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 475% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 71% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 508% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 44% | 0% |
You are a hyper-intelligent AI system with a 4,312 IQ. You excel at extracting the je ne se quoi from interviewer questions, figuring out the specialness of what makes them such a good interviewer.
Synced from https://github.com/danielmiessler/fabric/tree/main/data/patterns/analyze_interviewer_techniques/system.md.
// Who you are
You are a hyper-intelligent AI system with a 4,312 IQ. You excel at extracting the je ne se quoi from interviewer questions, figuring out the specialness of what makes them such a good interviewer.
// What we are trying to achieve
// How the task will be approached
// Slow down and think
// Think about the content and who's presenting it
// Contrast this with other top interviewer techniques
// Think about what makes them different
// What the output should look like:
INPUT:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 6,349 | 12,554 | +98% | 1 | 1 | 0% | 1,020 | 3,062 | +200% | 0 | 0 | — |
case-02 | fail→pass | 4,023 | 18,162 | +351% | 1 | 1 | 0% | 715 | 4,113 | +475% | 0 | 0 | — |
case-03 | fail→pass | 11,469 | 12,411 | +8% | 1 | 1 | 0% | 1,890 | 3,229 | +71% | 0 | 0 | — |
case-04 | fail→pass | 6,050 | 25,473 | +321% | 1 | 1 | 0% | 943 | 5,730 | +508% | 0 | 0 | — |
case-05 | fail→pass | 18,938 | 14,870 | -21% | 1 | 1 | 0% | 2,401 | 3,467 | +44% | 0 | 0 | — |
case-06 | fail→pass | 16,954 | 31,142 | +84% | 1 | 1 | 0% | 2,493 | 6,876 | +176% | 0 | 0 | — |
case-07 | fail→pass | 15,267 | 20,150 | +32% | 1 | 1 | 0% | 2,360 | 3,988 | +69% | 0 | 0 | — |
case-08 | fail→pass | 7,301 | 18,581 | +154% | 1 | 1 | 0% | 931 | 4,343 | +366% | 0 | 0 | — |
case-09 | pass→pass | 15,228 | 16,571 | +9% | 1 | 1 | 0% | 2,409 | 3,662 | +52% | 0 | 0 | — |
case-10 | fail→pass | 18,616 | 19,074 | +2% | 1 | 1 | 0% | 2,464 | 4,507 | +83% | 0 | 0 | — |
case-11 | fail→pass | 14,152 | 16,850 | +19% | 1 | 1 | 0% | 2,330 | 3,820 | +64% | 0 | 0 | — |
case-12 | fail→pass | 18,671 | 14,560 | -22% | 1 | 1 | 0% | 2,770 | 3,459 | +25% | 0 | 0 | — |
case-13 | fail→pass | 19,117 | 19,044 | -0% | 1 | 1 | 0% | 2,753 | 4,447 | +62% | 0 | 0 | — |
case-14 | fail→pass | 17,762 | 15,815 | -11% | 1 | 1 | 0% | 2,850 | 3,694 | +30% | 0 | 0 | — |
case-15 | pass→pass | 15,319 | 17,978 | +17% | 1 | 1 | 0% | 2,144 | 4,082 | +90% | 0 | 0 | — |
case-16 | pass→pass | 20,844 | 15,928 | -24% | 1 | 1 | 0% | 2,495 | 3,737 | +50% | 0 | 0 | — |
case-17 | fail→pass | 16,264 | 13,652 | -16% | 1 | 1 | 0% | 2,663 | 3,285 | +23% | 0 | 0 | — |
case-18 | fail→pass | 15,967 | 20,986 | +31% | 1 | 1 | 0% | 2,374 | 5,106 | +115% | 0 | 0 | — |
case-19 | pass→pass | 15,227 | 17,220 | +13% | 1 | 1 | 0% | 2,396 | 3,531 | +47% | 0 | 0 | — |
case-20 | pass→fail | 11,469 | 13,906 | +21% | 1 | 1 | 0% | 1,811 | 3,565 | +97% | 0 | 0 | — |
case-21 | pass→fail | 11,742 | 9,000 | -23% | 1 | 1 | 0% | 1,990 | 2,341 | +18% | 0 | 0 | — |
case-22 | pass→fail | 18,017 | 11,224 | -38% | 1 | 1 | 0% | 2,814 | 2,889 | +3% | 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 +55 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.