Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Validate Rust work after substantial Rust code changes by running `just check`, `just clippy`, then `just format`. Use before the final response after a significant Rust implementation or cleanup task; batch edits first instead of running after every small change.
.claude/skills/macro-inc-rust-validation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | -56% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -44% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -35% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -47% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 16% | 0% |
Use this skill when you have changed Rust code or Rust build inputs and are nearing the end of a substantial implementation, refactor, or bug-fix task.
Do not run these commands after every individual edit. Batch related edits first, then validate once at the end of a major work pass or before handing the task back to the user.
Run validation from the repository root. If the root justfile or one of the required recipes is missing, stop and tell the user what prevented validation instead of inventing alternate commands.
Run these commands one at a time, in this exact order:
bashjust rust-check just clippy just format just hakari
If just rust-check or just clippy fails, do not continue to later commands. Inspect the failure, fix it if it is in scope, then rerun the full sequence starting at just rust-check.
If just format changes files, review the resulting changes before the final response.
Briefly report whether validation passed and list the commands that ran. If validation was skipped because no Rust code was changed, or because the user explicitly asked to skip it, say so.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 7,176 | 5,916 | -18% | 1 | 1 | 0% | 423 | 488 | +15% | 0 | 0 | — |
case-02 | fail→fail | 5,102 | 6,304 | +24% | 1 | 1 | 0% | 196 | 502 | +156% | 0 | 0 | — |
case-03 | fail→fail | 9,251 | 7,702 | -17% | 1 | 1 | 0% | 466 | 565 | +21% | 0 | 0 | — |
case-04 | pass→pass | 8,042 | 5,061 | -37% | 1 | 1 | 0% | 1,126 | 929 | -17% | 0 | 0 | — |
case-05 | fail→pass | 11,167 | 3,585 | -68% | 1 | 1 | 0% | 1,638 | 725 | -56% | 0 | 0 | — |
case-06 | fail→pass | 9,193 | 3,687 | -60% | 1 | 1 | 0% | 1,391 | 781 | -44% | 0 | 0 | — |
case-07 | fail→pass | 14,356 | 4,778 | -67% | 1 | 1 | 0% | 1,309 | 848 | -35% | 0 | 0 | — |
case-08 | pass→pass | 9,569 | 3,323 | -65% | 1 | 1 | 0% | 1,184 | 737 | -38% | 0 | 0 | — |
case-09 | fail→fail | 8,497 | 4,510 | -47% | 1 | 1 | 0% | 1,227 | 1,013 | -17% | 0 | 0 | — |
case-10 | fail→pass | 14,654 | 6,099 | -58% | 1 | 1 | 0% | 1,713 | 907 | -47% | 0 | 0 | — |
case-11 | pass→pass | 10,808 | 3,750 | -65% | 1 | 1 | 0% | 1,628 | 793 | -51% | 0 | 0 | — |
case-12 | fail→pass | 5,133 | 4,296 | -16% | 1 | 1 | 0% | 566 | 658 | +16% | 0 | 0 | — |
case-13 | fail→fail | 12,777 | 5,219 | -59% | 1 | 1 | 0% | 2,019 | 840 | -58% | 0 | 0 | — |
case-14 | fail→pass | 7,479 | 3,261 | -56% | 1 | 1 | 0% | 932 | 664 | -29% | 0 | 0 | — |
case-15 | fail→pass | 9,300 | 4,059 | -56% | 1 | 1 | 0% | 1,432 | 836 | -42% | 0 | 0 | — |
case-16 | fail→pass | 9,487 | 3,756 | -60% | 1 | 1 | 0% | 1,273 | 806 | -37% | 0 | 0 | — |
case-17 | fail→pass | 11,845 | 5,756 | -51% | 1 | 1 | 0% | 1,740 | 976 | -44% | 0 | 0 | — |
case-18 | fail→fail | 8,736 | 3,802 | -56% | 1 | 1 | 0% | 1,328 | 729 | -45% | 0 | 0 | — |
case-19 | fail→pass | 20,861 | 4,467 | -79% | 1 | 1 | 0% | 1,343 | 713 | -47% | 0 | 0 | — |
case-20 | pass→pass | 9,646 | 8,617 | -11% | 1 | 1 | 0% | 1,359 | 848 | -38% | 0 | 0 | — |
case-21 | pass→pass | 13,186 | 21,535 | +63% | 1 | 1 | 0% | 2,196 | 1,986 | -10% | 0 | 0 | — |
case-22 | pass→pass | 6,147 | 5,785 | -6% | 1 | 1 | 0% | 952 | 1,117 | +17% | 0 | 0 | — |
case-23 | pass→pass | 4,110 | 4,736 | +15% | 1 | 1 | 0% | 677 | 908 | +34% | 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, and 20 counted toward the lift figure. The other 3 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 +43 percentage points is the difference between those two pass rates over the 20 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.