Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Adds developer-authored annotations to the gauntlet knowledge base. Use when capturing tribal knowledge or rationale not visible in code.
.claude/skills/athola-curate/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -62% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -56% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -27% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -61% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -63% | 0% |
Add developer-authored annotations to the knowledge base.
gauntlet:extract)gauntlet:gauntlet-curate).gauntlet/annotations/<slug>.yaml.gauntlet/annotations/<slug>.yamlafter the skill completes, where <slug> is derived from the annotated module and concept
concept field (keyinsight or rule) and a why field (rationale/history/context)
in future challenges, indicating it is on the discovery path for gauntlet:extract
same slug (checked before saving)
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 13,466 | 6,413 | -52% | 1 | 1 | 0% | 2,197 | 829 | -62% | 0 | 0 | — |
case-02 | fail→pass | 12,393 | 3,380 | -73% | 1 | 1 | 0% | 1,931 | 846 | -56% | 0 | 0 | — |
case-03 | fail→pass | 13,061 | 7,193 | -45% | 1 | 1 | 0% | 2,081 | 1,517 | -27% | 0 | 0 | — |
case-04 | fail→pass | 13,768 | 4,077 | -70% | 1 | 1 | 0% | 2,361 | 930 | -61% | 0 | 0 | — |
case-05 | fail→pass | 13,390 | 3,333 | -75% | 1 | 1 | 0% | 2,176 | 807 | -63% | 0 | 0 | — |
case-06 | fail→pass | 10,958 | 7,601 | -31% | 1 | 1 | 0% | 2,117 | 1,593 | -25% | 0 | 0 | — |
case-07 | fail→fail | 11,091 | 3,215 | -71% | 1 | 1 | 0% | 1,976 | 635 | -68% | 0 | 0 | — |
case-08 | fail→pass | 7,494 | 3,719 | -50% | 1 | 1 | 0% | 1,382 | 823 | -40% | 0 | 0 | — |
case-09 | fail→pass | 10,864 | 4,440 | -59% | 1 | 1 | 0% | 1,970 | 1,031 | -48% | 0 | 0 | — |
case-10 | pass→pass | 8,509 | 2,338 | -73% | 1 | 1 | 0% | 1,334 | 627 | -53% | 0 | 0 | — |
case-11 | fail→pass | 10,297 | 2,515 | -76% | 1 | 1 | 0% | 1,676 | 685 | -59% | 0 | 0 | — |
case-12 | fail→pass | 12,396 | 3,902 | -69% | 1 | 1 | 0% | 2,022 | 993 | -51% | 0 | 0 | — |
case-13 | pass→pass | 20,877 | 2,746 | -87% | 1 | 1 | 0% | 1,699 | 776 | -54% | 0 | 0 | — |
case-14 | pass→pass | 4,509 | 1,846 | -59% | 1 | 1 | 0% | 691 | 536 | -22% | 0 | 0 | — |
case-15 | fail→pass | 8,216 | 2,391 | -71% | 1 | 1 | 0% | 1,210 | 704 | -42% | 0 | 0 | — |
case-16 | fail→pass | 4,017 | 2,354 | -41% | 1 | 1 | 0% | 635 | 605 | -5% | 0 | 0 | — |
case-17 | fail→pass | 11,091 | 3,566 | -68% | 1 | 1 | 0% | 1,807 | 828 | -54% | 0 | 0 | — |
case-18 | fail→pass | 10,655 | 8,727 | -18% | 1 | 1 | 0% | 1,851 | 1,706 | -8% | 0 | 0 | — |
case-19 | fail→pass | 20,417 | 1,703 | -92% | 1 | 1 | 0% | 1,481 | 488 | -67% | 0 | 0 | — |
case-20 | fail→pass | 12,640 | 7,784 | -38% | 1 | 1 | 0% | 2,018 | 1,702 | -16% | 0 | 0 | — |
case-21 | pass→pass | 7,720 | 5,411 | -30% | 1 | 1 | 0% | 1,200 | 1,254 | +5% | 0 | 0 | — |
case-22 | fail→pass | 26,193 | 4,720 | -82% | 1 | 1 | 0% | 2,275 | 1,139 | -50% | 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. The headline lift of +77 percentage points is the difference between those two pass rates over the 22 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.