Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Strategy: Klein pre-mortem — assume the artifact has failed, then retrospect plausible causes. Generates rapid failure scenario catalog.
.claude/skills/yogsoth-ai-prospective-hindsight/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -1% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 18% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 27% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -52% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -19% | 0% |
Assume the artifact has already failed. Retrospect: what went wrong?
| Parameter | S | M | L | |---|---|---|---| | Failure scenarios generated | 8 | 20 | 40 | | Independent perspectives | 2 | 4 | 6 | | Screening threshold (severity) | 7 | 5 | 3 |
premortem-facilitation → failure-mode-extraction
→ severity-scoring (rapid screen)
→ [high-severity items] → premortem-to-fmea-pipeline<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| Tactic | When to use | | --- | --- | | failure-chain-tracing | Tactic: Trace upstream causes and downstream effects of each failure mode. Builds multi-level cause-mode-effect chains for systemic understanding. | | premortem-to-fmea-pipeline | Tactic: Pre-mortem rapid screening feeds high-risk items into full FMEA analysis. Bridges fast intuitive generation with systematic structured analysis. |
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | failure-mode-extraction | Extract structured failure mode list from raw scenarios or artifact analysis. Produces standardized failure mode records. | | premortem-facilitation | Execute Klein pre-mortem protocol — assume failure has occurred, generate plausible failure scenarios through prospective hindsight. | | severity-scoring | Rate failure mode severity 1-10 based on end-effect impact. Follows AIAG-VDA severity scale calibrated for research artifacts. |
<!-- 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-01 | fail→fail | 41,185 | 52,070 | +26% | 1 | 1 | 0% | 5,916 | 8,312 | +41% | 0 | 0 | — |
case-02 | fail→pass | 41,020 | 39,728 | -3% | 1 | 1 | 0% | 6,397 | 6,361 | -1% | 0 | 0 | — |
case-03 | fail→pass | 25,354 | 27,690 | +9% | 1 | 1 | 0% | 3,548 | 4,169 | +18% | 0 | 0 | — |
case-04 | pass→fail | 22,283 | 32,438 | +46% | 1 | 1 | 0% | 3,830 | 4,954 | +29% | 0 | 0 | — |
case-05 | fail→fail | 10,557 | 27,712 | +162% | 1 | 1 | 0% | 548 | 4,308 | +686% | 0 | 0 | — |
case-06 | fail→pass | 24,503 | 3,141 | -87% | 1 | 1 | 0% | 693 | 883 | +27% | 0 | 0 | — |
case-07 | fail→pass | 13,366 | 4,863 | -64% | 1 | 1 | 0% | 2,175 | 1,035 | -52% | 0 | 0 | — |
case-08 | fail→pass | 22,863 | 7,559 | -67% | 1 | 1 | 0% | 1,034 | 837 | -19% | 0 | 0 | — |
case-09 | fail→pass | 20,195 | 19,623 | -3% | 1 | 1 | 0% | 3,565 | 2,967 | -17% | 0 | 0 | — |
case-10 | fail→pass | 11,230 | 7,670 | -32% | 1 | 1 | 0% | 993 | 1,021 | +3% | 0 | 0 | — |
case-11 | pass→fail | 14,135 | 9,983 | -29% | 1 | 1 | 0% | 1,504 | 1,154 | -23% | 0 | 0 | — |
case-12 | pass→pass | 15,336 | 14,325 | -7% | 1 | 1 | 0% | 2,416 | 2,161 | -11% | 0 | 0 | — |
case-13 | pass→pass | 23,546 | 16,499 | -30% | 1 | 1 | 0% | 2,933 | 2,370 | -19% | 0 | 0 | — |
case-14 | pass→pass | 20,373 | 10,084 | -51% | 1 | 1 | 0% | 2,190 | 1,173 | -46% | 0 | 0 | — |
case-15 | pass→pass | 11,856 | 9,489 | -20% | 1 | 1 | 0% | 1,789 | 1,108 | -38% | 0 | 0 | — |
case-16 | fail→pass | 16,116 | 4,528 | -72% | 1 | 1 | 0% | 1,661 | 1,267 | -24% | 0 | 0 | — |
case-17 | fail→pass | 20,992 | 25,066 | +19% | 1 | 1 | 0% | 3,429 | 3,974 | +16% | 0 | 0 | — |
case-18 | fail→pass | 38,680 | 28,610 | -26% | 1 | 1 | 0% | 5,982 | 5,378 | -10% | 0 | 0 | — |
case-19 | fail→pass | 36,799 | 49,925 | +36% | 1 | 1 | 0% | 5,681 | 7,716 | +36% | 0 | 0 | — |
case-20 | fail→pass | 18,351 | 8,643 | -53% | 1 | 1 | 0% | 3,118 | 1,550 | -50% | 0 | 0 | — |
case-21 | fail→pass | 20,013 | 17,897 | -11% | 1 | 1 | 0% | 2,818 | 2,958 | +5% | 0 | 0 | — |
case-22 | pass→pass | 26,640 | 23,468 | -12% | 1 | 1 | 0% | 1,815 | 2,293 | +26% | 0 | 0 | — |
case-23 | pass→pass | 13,361 | 4,904 | -63% | 1 | 1 | 0% | 1,259 | 1,222 | -3% | 0 | 0 | — |
case-24 | pass→pass | 14,785 | 10,552 | -29% | 1 | 1 | 0% | 1,711 | 2,080 | +22% | 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. 24 cases were attempted, and 22 counted toward the lift figure. The other 2 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 +46 percentage points is the difference between those two pass rates over the 22 comparable cases. 2 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.