Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Master enterprise-grade Scala development with functional programming, distributed systems, and big data processing. Expert in Apache Pekko, Akka, Spark, ZIO/Cats Effect, and reactive architectures.
.claude/skills/davila7-scala-pro/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-10 | ✗→✓ | ▲ Improved | 85% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 21% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | 8% | 0% |
| case-06 | ✓→✓ | ＝ Same ✓ | 28% | 0% |
| case-08 | ✓→✓ | ＝ Same ✓ | 6% | 0% |
resources/implementation-playbook.md.You are an elite Scala engineer specializing in enterprise-grade functional programming and distributed systems.
given/using clauses for context functions, and metaprogramming with inline and macrosFuture composition, parallel collections, and principled concurrency using effect systems over manual thread managementEither, Validated, and Ior from the Cats library, or using ZIO's integrated error channelDeliver robust, maintainable, and performant Scala solutions that scale to millions of users.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | pass→pass | 16,467 | 13,853 | -16% | 1 | 1 | 0% | 3,307 | 3,996 | +21% | 0 | 0 | — |
case-11 | fail→fail | 8,437 | 5,890 | -30% | 1 | 1 | 0% | 1,575 | 2,079 | +32% | 0 | 0 | — |
case-01 | fail→fail | 20,818 | 18,249 | -12% | 1 | 1 | 0% | 4,274 | 5,047 | +18% | 0 | 0 | — |
case-02 | fail→fail | 21,736 | 18,430 | -15% | 1 | 1 | 0% | 4,157 | 4,730 | +14% | 0 | 0 | — |
case-03 | fail→fail | 26,065 | 18,420 | -29% | 1 | 1 | 0% | 5,686 | 5,025 | -12% | 0 | 0 | — |
case-05 | pass→pass | 18,698 | 15,000 | -20% | 1 | 1 | 0% | 3,586 | 3,888 | +8% | 0 | 0 | — |
case-06 | pass→pass | 17,408 | 15,031 | -14% | 1 | 1 | 0% | 3,166 | 4,060 | +28% | 0 | 0 | — |
case-07 | fail→fail | 14,776 | 8,777 | -41% | 1 | 1 | 0% | 2,410 | 2,669 | +11% | 0 | 0 | — |
case-08 | pass→pass | 16,467 | 12,021 | -27% | 1 | 1 | 0% | 2,932 | 3,107 | +6% | 0 | 0 | — |
case-09 | fail→fail | 18,732 | 16,179 | -14% | 1 | 1 | 0% | 3,670 | 4,369 | +19% | 0 | 0 | — |
case-10 | fail→pass | 8,430 | 8,746 | +4% | 1 | 1 | 0% | 1,508 | 2,786 | +85% | 0 | 0 | — |
case-12 | fail→fail | 5,917 | 6,971 | +18% | 1 | 1 | 0% | 1,038 | 2,150 | +107% | 0 | 0 | — |
case-13 | pass→pass | 12,164 | 11,528 | -5% | 1 | 1 | 0% | 2,384 | 3,330 | +40% | 0 | 0 | — |
case-14 | fail→fail | 12,713 | 9,981 | -21% | 1 | 1 | 0% | 2,279 | 2,876 | +26% | 0 | 0 | — |
case-15 | fail→fail | 10,992 | 8,787 | -20% | 1 | 1 | 0% | 2,084 | 2,680 | +29% | 0 | 0 | — |
case-16 | pass→pass | 11,054 | 9,437 | -15% | 1 | 1 | 0% | 2,049 | 2,719 | +33% | 0 | 0 | — |
case-17 | fail→fail | 13,089 | 10,134 | -23% | 1 | 1 | 0% | 2,563 | 3,112 | +21% | 0 | 0 | — |
case-18 | fail→fail | 13,970 | 12,707 | -9% | 1 | 1 | 0% | 2,479 | 3,443 | +39% | 0 | 0 | — |
case-19 | fail→fail | 11,809 | 8,404 | -29% | 1 | 1 | 0% | 2,114 | 2,626 | +24% | 0 | 0 | — |
case-20 | fail→fail | 4,503 | 4,037 | -10% | 1 | 1 | 0% | 821 | 1,816 | +121% | 0 | 0 | — |
case-21 | fail→fail | 17,373 | 6,587 | -62% | 1 | 1 | 0% | 3,072 | 2,209 | -28% | 0 | 0 | — |
case-22 | pass→pass | 11,707 | 8,766 | -25% | 1 | 1 | 0% | 2,176 | 2,682 | +23% | 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 +5 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.