Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Generate comprehensive performance reports with insights and actionable recommendations.
.claude/skills/holaboss-ai-performance-reporter/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -4% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 34% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 61% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 43% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -34% | 0% |
Report like an analyst the team actually wants to hear from: turn raw metrics into a short narrative that says what happened, why it matters, and what to do next. Numbers are evidence; the insight is the product.
Use Performance Reporter to interpret social/content metrics over time: engagement, reach, growth, and platform comparisons, rolled into a readable report with recommendations. If the task is profiling who the audience is, use Audience Analyst instead.
Flag data quality issues (missing days, platform changes, anomalies) rather than reporting on top of broken inputs.
Structure the report for a busy reader:
Lead with the summary. Keep tables tight and the prose tighter.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 19,494 | 21,169 | +9% | 1 | 1 | 0% | 3,644 | 3,484 | -4% | 0 | 0 | — |
case-02 | pass→fail | 14,159 | 9,581 | -32% | 1 | 1 | 0% | 2,412 | 1,734 | -28% | 0 | 0 | — |
case-21 | pass→pass | 18,383 | 20,403 | +11% | 1 | 1 | 0% | 2,826 | 3,337 | +18% | 0 | 0 | — |
case-03 | pass→pass | 19,547 | 19,495 | -0% | 1 | 1 | 0% | 2,850 | 3,162 | +11% | 0 | 0 | — |
case-04 | pass→pass | 17,769 | 17,611 | -1% | 1 | 1 | 0% | 2,770 | 2,959 | +7% | 0 | 0 | — |
case-05 | fail→pass | 13,287 | 13,487 | +2% | 1 | 1 | 0% | 1,890 | 2,536 | +34% | 0 | 0 | — |
case-06 | fail→pass | 11,306 | 16,383 | +45% | 1 | 1 | 0% | 1,650 | 2,662 | +61% | 0 | 0 | — |
case-07 | pass→pass | 14,733 | 17,267 | +17% | 1 | 1 | 0% | 2,254 | 3,246 | +44% | 0 | 0 | — |
case-08 | pass→pass | 15,000 | 12,333 | -18% | 1 | 1 | 0% | 2,099 | 2,277 | +8% | 0 | 0 | — |
case-09 | fail→pass | 19,228 | 20,612 | +7% | 1 | 1 | 0% | 2,522 | 3,618 | +43% | 0 | 0 | — |
case-10 | pass→pass | 21,147 | 16,464 | -22% | 1 | 1 | 0% | 2,989 | 2,999 | +0% | 0 | 0 | — |
case-11 | fail→pass | 13,984 | 6,634 | -53% | 1 | 1 | 0% | 2,000 | 1,320 | -34% | 0 | 0 | — |
case-12 | fail→pass | 12,355 | 7,384 | -40% | 1 | 1 | 0% | 2,017 | 1,761 | -13% | 0 | 0 | — |
case-13 | pass→pass | 16,548 | 13,293 | -20% | 1 | 1 | 0% | 2,534 | 2,419 | -5% | 0 | 0 | — |
case-14 | pass→pass | 18,011 | 21,121 | +17% | 1 | 1 | 0% | 2,631 | 3,441 | +31% | 0 | 0 | — |
case-15 | pass→pass | 21,370 | 15,632 | -27% | 1 | 1 | 0% | 2,265 | 2,573 | +14% | 0 | 0 | — |
case-16 | pass→pass | 12,616 | 5,742 | -54% | 1 | 1 | 0% | 1,920 | 1,349 | -30% | 0 | 0 | — |
case-17 | pass→pass | 9,079 | 11,715 | +29% | 1 | 1 | 0% | 1,244 | 2,044 | +64% | 0 | 0 | — |
case-18 | pass→pass | 27,284 | 15,611 | -43% | 1 | 1 | 0% | 3,331 | 3,147 | -6% | 0 | 0 | — |
case-19 | pass→pass | 19,990 | 18,296 | -8% | 1 | 1 | 0% | 2,807 | 2,992 | +7% | 0 | 0 | — |
case-20 | fail→pass | 19,260 | 19,796 | +3% | 1 | 1 | 0% | 2,906 | 3,171 | +9% | 0 | 0 | — |
case-22 | pass→pass | 19,207 | 16,169 | -16% | 1 | 1 | 0% | 2,962 | 3,173 | +7% | 0 | 0 | — |
case-23 | pass→pass | 21,614 | 11,924 | -45% | 1 | 1 | 0% | 3,237 | 2,471 | -24% | 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. 23 cases were attempted. The headline lift of +26 percentage points is the difference between those two pass rates over the 23 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.