Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Evaluate a specific portfolio's performance metrics and vulnerabilities under a given scenario.
.claude/skills/yogsoth-ai-portfolio-evaluation-per-scenario/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 132% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 54% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 37% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 54% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 137% | 0% |
Assess how a specific portfolio performs under a given future scenario, identifying performance metrics and vulnerabilities.
Spawns a subagent that evaluates each portfolio member's performance under scenario conditions and aggregates into portfolio-level metrics.
Per-scenario evaluation requires careful reasoning about how each scenario's conditions affect each portfolio member differently. This analytical work is repeated per scenario and benefits from consistent, focused execution.
Output must include quantified performance metrics for the portfolio under the scenario and identification of any members that become vulnerable or fail under scenario conditions.
<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | spawn-agent | Spawn a customized CC subagent with full MCP tool access. Used by SOPs that declare execution: subagent. |
<!-- END available-tables (generated) -->
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | fail→fail | 35,681 | 111,611 | +213% | 1 | 1 | 0% | 5,402 | 762 | -86% | 0 | 0 | — |
case-01 | pass→pass | 38,452 | 61,750 | +61% | 1 | 1 | 0% | 2,534 | 5,995 | +137% | 0 | 0 | — |
case-03 | fail→fail | 27,438 | 40,593 | +48% | 1 | 1 | 0% | 3,815 | 7,116 | +87% | 0 | 0 | — |
case-04 | fail→pass | 23,110 | 48,846 | +111% | 1 | 1 | 0% | 3,359 | 7,789 | +132% | 0 | 0 | — |
case-05 | fail→pass | 27,858 | 34,659 | +24% | 1 | 1 | 0% | 3,611 | 5,562 | +54% | 0 | 0 | — |
case-06 | fail→fail | 28,601 | 39,232 | +37% | 1 | 1 | 0% | 3,648 | 6,512 | +79% | 0 | 0 | — |
case-07 | pass→pass | 24,430 | 44,652 | +83% | 1 | 1 | 0% | 4,472 | 7,380 | +65% | 0 | 0 | — |
case-08 | fail→fail | 31,137 | 46,225 | +48% | 1 | 1 | 0% | 4,650 | 8,423 | +81% | 0 | 0 | — |
case-09 | pass→pass | 23,117 | 33,387 | +44% | 1 | 1 | 0% | 3,625 | 5,415 | +49% | 0 | 0 | — |
case-10 | fail→fail | 23,837 | 37,072 | +56% | 1 | 1 | 0% | 2,774 | 5,070 | +83% | 0 | 0 | — |
case-11 | pass→pass | 21,468 | 38,078 | +77% | 1 | 1 | 0% | 2,856 | 6,427 | +125% | 0 | 0 | — |
case-12 | fail→fail | 36,469 | 35,647 | -2% | 1 | 1 | 0% | 3,844 | 5,551 | +44% | 0 | 0 | — |
case-13 | fail→pass | 30,841 | 28,882 | -6% | 1 | 1 | 0% | 3,233 | 4,424 | +37% | 0 | 0 | — |
case-14 | fail→fail | 28,629 | 50,215 | +75% | 1 | 1 | 0% | 3,780 | 8,435 | +123% | 0 | 0 | — |
case-15 | fail→pass | 26,235 | 27,447 | +5% | 1 | 1 | 0% | 3,180 | 4,900 | +54% | 0 | 0 | — |
case-16 | fail→fail | 21,889 | 42,023 | +92% | 1 | 1 | 0% | 3,615 | 7,684 | +113% | 0 | 0 | — |
case-17 | fail→pass | 24,988 | 45,455 | +82% | 1 | 1 | 0% | 3,557 | 8,422 | +137% | 0 | 0 | — |
case-18 | fail→fail | 22,563 | 34,301 | +52% | 1 | 1 | 0% | 3,278 | 6,358 | +94% | 0 | 0 | — |
case-19 | fail→pass | 28,532 | 35,177 | +23% | 1 | 1 | 0% | 2,940 | 7,008 | +138% | 0 | 0 | — |
case-20 | pass→pass | 23,718 | 33,884 | +43% | 1 | 1 | 0% | 4,185 | 6,472 | +55% | 0 | 0 | — |
case-21 | pass→pass | 21,965 | 26,263 | +20% | 1 | 1 | 0% | 3,446 | 4,218 | +22% | 0 | 0 | — |
case-22 | pass→pass | 14,374 | 33,098 | +130% | 1 | 1 | 0% | 1,385 | 2,897 | +109% | 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, and 21 counted toward the lift figure. The other 1 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +27 percentage points is the difference between those two pass rates over the 21 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.