Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when a framed and root-caused problem needs its assumptions made explicit before design or planning begins — explicit Assumptions, Unknowns, and a Validation plan. This should trigger when an issue's Assumption Analysis point of view needs evaluation, or when a maintainer directly asks to surface hidden assumptions and unknowns before committing to an approach. Part of Plinth Toolkit
.claude/skills/jabrena-023-assumption-analysis/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | 53% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 912% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -10% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 119% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 99% | 0% |
Guide production of explicit Assumptions, a list of Unknowns, and a Validation plan for a problem under exploration. This is an interactive SKILL.
What is covered in this Skill?
024-context-mapping and the remaining Functional Specification lensesMake assumptions and unknowns explicit before they become undiscussed risk. When this technique is orchestrated by another workflow, the orchestrator owns clarifying-question sequencing; when applied standalone, ask directly.
references/023-assumption-analysis.md before applying Assumption Analysis guidanceRead references/023-assumption-analysis.md, then review the problem frame and root-cause findings for implicit beliefs.
State each assumption as a falsifiable claim believed true but not yet verified.
List facts that are not yet known either way, distinct from assumptions.
Rank assumptions and unknowns by impact if wrong and by current confidence, prioritizing high-impact, low-confidence items.
Name how and when each high-priority assumption or unknown will be validated.
Report the Assumptions, Unknowns, and Validation plan, and flag any item left open pending a clarifying answer.
For detailed guidance, examples, and constraints, see references/023-assumption-analysis.md.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-22 | pass→fail | 18,679 | 16,466 | -12% | 1 | 1 | 0% | 2,575 | 3,064 | +19% | 0 | 0 | — |
case-01 | fail→fail | 23,332 | 18,245 | -22% | 1 | 1 | 0% | 3,394 | 3,671 | +8% | 0 | 0 | — |
case-02 | fail→fail | 23,864 | 23,059 | -3% | 1 | 1 | 0% | 3,612 | 4,278 | +18% | 0 | 0 | — |
case-03 | fail→fail | 11,973 | 22,994 | +92% | 1 | 1 | 0% | 1,831 | 4,192 | +129% | 0 | 0 | — |
case-04 | fail→fail | 8,538 | 11,800 | +38% | 1 | 1 | 0% | 1,334 | 2,520 | +89% | 0 | 0 | — |
case-05 | fail→pass | 11,172 | 11,630 | +4% | 1 | 1 | 0% | 1,613 | 2,461 | +53% | 0 | 0 | — |
case-20 | pass→fail | 7,700 | 17,925 | +133% | 1 | 1 | 0% | 1,403 | 3,494 | +149% | 0 | 0 | — |
case-06 | fail→fail | 3,505 | 5,350 | +53% | 1 | 1 | 0% | 632 | 1,490 | +136% | 0 | 0 | — |
case-07 | fail→pass | 1,420 | 10,408 | +633% | 1 | 1 | 0% | 218 | 2,206 | +912% | 0 | 0 | — |
case-08 | fail→pass | 19,123 | 14,027 | -27% | 1 | 1 | 0% | 2,964 | 2,659 | -10% | 0 | 0 | — |
case-09 | pass→pass | 14,152 | 7,950 | -44% | 1 | 1 | 0% | 1,986 | 1,729 | -13% | 0 | 0 | — |
case-10 | fail→pass | 8,442 | 12,543 | +49% | 1 | 1 | 0% | 1,188 | 2,600 | +119% | 0 | 0 | — |
case-21 | pass→fail | 17,666 | 23,825 | +35% | 1 | 1 | 0% | 2,783 | 4,169 | +50% | 0 | 0 | — |
case-11 | pass→pass | 10,382 | 16,291 | +57% | 1 | 1 | 0% | 1,668 | 3,079 | +85% | 0 | 0 | — |
case-12 | fail→pass | 8,657 | 10,267 | +19% | 1 | 1 | 0% | 1,162 | 2,311 | +99% | 0 | 0 | — |
case-13 | fail→pass | 16,386 | 13,250 | -19% | 1 | 1 | 0% | 2,350 | 2,644 | +13% | 0 | 0 | — |
case-14 | fail→fail | 2,263 | 13,125 | +480% | 1 | 1 | 0% | 363 | 2,701 | +644% | 0 | 0 | — |
case-15 | fail→pass | 9,164 | 6,024 | -34% | 1 | 1 | 0% | 1,331 | 1,417 | +6% | 0 | 0 | — |
case-16 | fail→fail | 21,479 | 21,647 | +1% | 1 | 1 | 0% | 3,202 | 3,968 | +24% | 0 | 0 | — |
case-17 | fail→pass | 1,910 | 19,541 | +923% | 1 | 1 | 0% | 285 | 2,892 | +915% | 0 | 0 | — |
case-18 | fail→pass | 3,453 | 7,312 | +112% | 1 | 1 | 0% | 499 | 1,889 | +279% | 0 | 0 | — |
case-19 | pass→pass | 17,868 | 17,929 | +0% | 1 | 1 | 0% | 2,642 | 3,072 | +16% | 0 | 0 | — |
case-23 | pass→pass | 13,875 | 26,998 | +95% | 1 | 1 | 0% | 2,142 | 4,233 | +98% | 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. 4 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.