Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Enforce existence, reuse, mirror, and symmetry principles to keep new code minimal and consistent with surrounding code. Use when writing new code in an existing codebase, adding new features, refactoring, or making any code changes.
.claude/skills/tobihagemann-code-style/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -25% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -7% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -27% | 0% |
| case-03 | ✓→✓ | = Same ✓ | -37% | 0% |
| case-13 | ✓→✓ | = Same ✓ | 54% | 0% |
When writing new code in an existing codebase, work through these in order. When two pull in different directions, the earlier one wins.
fooB() parallel to fooA(), match naming, parameter order, and structure. Symmetry covers every name the new sibling makes asymmetric: identifiers, route paths, config keys, event names, environment variables, CLI flags, and file naming schemes. Rename the existing name by default; keep it as it stands only when it is a published contract whose consumers you cannot update, or cannot update in step with the rename. Treat deployment and registration as costs the change absorbs: updating every place the name is referenced or registered is part of the change.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 12,480 | 7,304 | -41% | 1 | 1 | 0% | 2,505 | 1,870 | -25% | 0 | 0 | — |
case-02 | fail→fail | 8,973 | 6,709 | -25% | 1 | 1 | 0% | 1,966 | 1,753 | -11% | 0 | 0 | — |
case-03 | pass→pass | 10,315 | 3,166 | -69% | 1 | 1 | 0% | 1,658 | 1,044 | -37% | 0 | 0 | — |
case-13 | pass→pass | 4,254 | 3,328 | -22% | 1 | 1 | 0% | 683 | 1,052 | +54% | 0 | 0 | — |
case-04 | pass→pass | 12,418 | 3,482 | -72% | 1 | 1 | 0% | 2,014 | 1,058 | -47% | 0 | 0 | — |
case-05 | pass→pass | 12,223 | 4,227 | -65% | 1 | 1 | 0% | 1,897 | 1,242 | -35% | 0 | 0 | — |
case-06 | pass→pass | 9,370 | 6,760 | -28% | 1 | 1 | 0% | 1,658 | 1,615 | -3% | 0 | 0 | — |
case-07 | pass→pass | 8,630 | 2,400 | -72% | 1 | 1 | 0% | 1,462 | 939 | -36% | 0 | 0 | — |
case-08 | pass→pass | 11,398 | 3,243 | -72% | 1 | 1 | 0% | 1,704 | 1,054 | -38% | 0 | 0 | — |
case-09 | pass→pass | 5,636 | 3,124 | -45% | 1 | 1 | 0% | 927 | 1,024 | +10% | 0 | 0 | — |
case-10 | pass→pass | 10,149 | 5,497 | -46% | 1 | 1 | 0% | 1,825 | 1,487 | -19% | 0 | 0 | — |
case-11 | fail→pass | 10,122 | 5,639 | -44% | 1 | 1 | 0% | 1,623 | 1,517 | -7% | 0 | 0 | — |
case-12 | fail→pass | 7,849 | 2,723 | -65% | 1 | 1 | 0% | 1,380 | 1,003 | -27% | 0 | 0 | — |
case-14 | pass→pass | 9,278 | 2,955 | -68% | 1 | 1 | 0% | 1,490 | 1,032 | -31% | 0 | 0 | — |
case-15 | pass→pass | 7,412 | 4,168 | -44% | 1 | 1 | 0% | 1,356 | 1,211 | -11% | 0 | 0 | — |
case-16 | pass→pass | 10,986 | 2,697 | -75% | 1 | 1 | 0% | 1,696 | 953 | -44% | 0 | 0 | — |
case-17 | pass→pass | 4,277 | 2,384 | -44% | 1 | 1 | 0% | 686 | 926 | +35% | 0 | 0 | — |
case-18 | pass→pass | 7,859 | 2,261 | -71% | 1 | 1 | 0% | 1,336 | 878 | -34% | 0 | 0 | — |
case-19 | fail→fail | 7,425 | 5,733 | -23% | 1 | 1 | 0% | 1,534 | 1,579 | +3% | 0 | 0 | — |
case-20 | pass→pass | 18,417 | 11,615 | -37% | 1 | 1 | 0% | 2,998 | 2,345 | -22% | 0 | 0 | — |
case-21 | pass→pass | 18,117 | 15,363 | -15% | 1 | 1 | 0% | 3,260 | 3,279 | +1% | 0 | 0 | — |
case-22 | pass→pass | 15,954 | 7,369 | -54% | 1 | 1 | 0% | 2,739 | 1,789 | -35% | 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 +14 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.