Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Applies TRIZ cross-domain analogical reasoning to find solutions from adjacent fields. Use when stuck on a problem and needing inventive perspectives.
.claude/skills/athola-triz/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-14 | ✗→✓ | ▲ Improved | 13% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 137% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | 7% | 0% |
| case-06 | ✓→✓ | ＝ Same ✓ | 63% | 0% |
| case-12 | ✓→✓ | ＝ Same ✓ | 57% | 0% |
tome channels)
Apply Altshuller's Theory of Inventive Problem Solving to find solutions from adjacent fields.
| Depth | Fields | Analysis | |-------|--------|----------| | light | 1 | Ideality and contradiction | | medium | 2 | Ideality, contradiction, field mapping | | deep | 3 | Adds principle suggestions and separation | | maximum | 5 | Adds distant fields and optional matrix lookup |
without the system existing. Ask what would make the system unnecessary while the function still happens.
worsens Y. For a physical contradiction (one parameter pulled toward two opposite values), apply separation in time, space, condition, or system/scale instead of compromise.
score.
TRIZ is the analogical method in the broader ideation catalog. For diverse, category-spanning ideation with rotation, see Skill(tome:ideate).
trade-offs to principles. It is a convenience mapping rather than part of the classical TRIZ Body of Knowledge, which is scoped to technological systems.
Altshuller's 39x39 engineering-parameter table. It has been frozen since 1985 and uses engineering, not software, parameters. Treat it as a cross-check, not the primary source.
empty-box convention, any of the 40 principles may apply.
principles as a divergence checklist and Ideality as a framing question. ARIZ, Substance-Field analysis, the 76 standard solutions, the Laws of Technical Systems Evolution (S-curves), and Function-Oriented Search (FOS) are out of scope here.
contradiction matrix (matriz.org).
across TRIZ steps; companion to AutoTRIZ.
NickScherbakov/Heinrich-The-Inventing-Machine (Apache-2.0); see src/tome/channels/triz_data/NOTICE for attribution and the exact vendored scope.
search begins.
worsens Y" (or a physical contradiction is named with a separation axis).
score is returned per active adjacent field, or the field is explicitly reported as yielding nothing.
is reported as "any of the 40 may apply", not as "no solution".
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 14,547 | 9,482 | -35% | 1 | 1 | 0% | 2,256 | 2,403 | +7% | 0 | 0 | — |
case-06 | pass→pass | 14,661 | 19,993 | +36% | 1 | 1 | 0% | 2,305 | 3,749 | +63% | 0 | 0 | — |
case-12 | pass→pass | 14,830 | 29,231 | +97% | 1 | 1 | 0% | 2,197 | 3,460 | +57% | 0 | 0 | — |
case-02 | pass→pass | 5,909 | 15,607 | +164% | 1 | 1 | 0% | 881 | 3,227 | +266% | 0 | 0 | — |
case-03 | pass→pass | 17,909 | 20,982 | +17% | 1 | 1 | 0% | 2,651 | 3,968 | +50% | 0 | 0 | — |
case-04 | pass→pass | 17,397 | 20,665 | +19% | 1 | 1 | 0% | 2,671 | 4,260 | +59% | 0 | 0 | — |
case-05 | pass→pass | 11,759 | 22,878 | +95% | 1 | 1 | 0% | 1,739 | 3,039 | +75% | 0 | 0 | — |
case-07 | pass→pass | 11,948 | 13,889 | +16% | 1 | 1 | 0% | 1,700 | 2,991 | +76% | 0 | 0 | — |
case-08 | pass→pass | 12,167 | 14,388 | +18% | 1 | 1 | 0% | 1,742 | 3,040 | +75% | 0 | 0 | — |
case-09 | pass→pass | 12,673 | 16,518 | +30% | 1 | 1 | 0% | 1,961 | 3,364 | +72% | 0 | 0 | — |
case-10 | pass→pass | 16,013 | 22,748 | +42% | 1 | 1 | 0% | 2,234 | 3,705 | +66% | 0 | 0 | — |
case-11 | pass→pass | 23,101 | 18,647 | -19% | 1 | 1 | 0% | 1,995 | 3,721 | +87% | 0 | 0 | — |
case-13 | pass→pass | 12,423 | 16,177 | +30% | 1 | 1 | 0% | 1,916 | 3,233 | +69% | 0 | 0 | — |
case-14 | fail→pass | 13,043 | 8,466 | -35% | 1 | 1 | 0% | 1,979 | 2,232 | +13% | 0 | 0 | — |
case-15 | pass→pass | 19,444 | 41,910 | +116% | 1 | 1 | 0% | 3,373 | 4,688 | +39% | 0 | 0 | — |
case-16 | pass→pass | 13,790 | 15,605 | +13% | 1 | 1 | 0% | 1,992 | 3,300 | +66% | 0 | 0 | — |
case-17 | pass→pass | 10,653 | 15,531 | +46% | 1 | 1 | 0% | 1,748 | 3,279 | +88% | 0 | 0 | — |
case-18 | pass→pass | 12,473 | 17,515 | +40% | 1 | 1 | 0% | 1,823 | 3,287 | +80% | 0 | 0 | — |
case-19 | fail→pass | 30,577 | 11,545 | -62% | 1 | 1 | 0% | 1,149 | 2,726 | +137% | 0 | 0 | — |
case-20 | pass→pass | 19,059 | 17,810 | -7% | 1 | 1 | 0% | 2,954 | 3,742 | +27% | 0 | 0 | — |
case-21 | pass→pass | 9,613 | 9,851 | +2% | 1 | 1 | 0% | 1,737 | 2,414 | +39% | 0 | 0 | — |
case-22 | pass→pass | 23,154 | 24,555 | +6% | 1 | 1 | 0% | 3,464 | 4,729 | +37% | 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 +9 percentage points is the difference between those two pass rates over the 21 comparable cases.
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.