Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Master Rust 1.75+ development with async runtime (Tokio/smol).
.claude/skills/dokhacgiakhoa-rust-pro/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 5% | 0% |
| case-04 | ✓→✗ | ▼ Worse | 21% | 0% |
| case-03 | ✓→✓ | = Same ✓ | 3% | 0% |
| case-05 | ✓→✓ | = Same ✓ | -8% | 0% |
> Goal: Write idiomatic, high-performance, and memory-safe Rust code following standard community practices (The Rust Way).
.clone() unless necessary. Use Arc<Mutex<T>> or RwLock<T> for shared state only when message passing (mpsc) is not viable.Result<T, E> with thiserror for libraries and anyhow for applications. Never use .unwrap() in production code; use .expect() with a context message or ? operator.tokio for general purpose apps. Use join_all for parallel execution of futures.New Type pattern to enforce validation at compile time.cargoclippy (Treat warnings as errors in CI)rustfmt#[test] and cargo test. Use mockall for mocking traits.textmy_crate/ ├── Cargo.toml ├── src/ │ ├── main.rs # Binary entry point │ ├── lib.rs # Library entry point │ ├── bin/ # Additional binaries │ ├── models/ # Data structures │ ├── error.rs # Central error definition │ └── utils.rs # Helper functions └── tests/ # Integration tests └── integration_test.rs
tokio, futuresaxum or actix-webserde, serde_jsonanyhow, thiserrortracing, tracing-subscriberconfig crate for environment managementString only when ownership is needed; prefer &str for function arguments.unsafe blocks unless absolutely necessary and documented with // SAFETY: comment explaining why it holds.Vec::with_capacity(n) if size is known.struct and enum definitions.main.rs or lib.rs.tests/.Anti-Patterns to Avoid:
Box<dyn Trait> (prefer generics with static dispatch).Result (always handle or propagate).| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 21,119 | 19,793 | -6% | 1 | 1 | 0% | 4,113 | 4,544 | +10% | 0 | 0 | — |
case-02 | fail→pass | 20,328 | 27,323 | +34% | 1 | 1 | 0% | 4,114 | 5,098 | +24% | 0 | 0 | — |
case-03 | pass→pass | 17,501 | 17,743 | +1% | 1 | 1 | 0% | 3,334 | 3,448 | +3% | 0 | 0 | — |
case-04 | pass→fail | 16,013 | 16,534 | +3% | 1 | 1 | 0% | 2,817 | 3,417 | +21% | 0 | 0 | — |
case-05 | pass→pass | 12,193 | 7,633 | -37% | 1 | 1 | 0% | 2,197 | 2,029 | -8% | 0 | 0 | — |
case-06 | pass→pass | 6,325 | 5,228 | -17% | 1 | 1 | 0% | 1,088 | 1,676 | +54% | 0 | 0 | — |
case-07 | pass→pass | 10,392 | 7,117 | -32% | 1 | 1 | 0% | 1,943 | 1,960 | +1% | 0 | 0 | — |
case-08 | pass→pass | 9,346 | 8,063 | -14% | 1 | 1 | 0% | 1,622 | 2,213 | +36% | 0 | 0 | — |
case-09 | pass→pass | 13,790 | 13,619 | -1% | 1 | 1 | 0% | 2,438 | 3,018 | +24% | 0 | 0 | — |
case-10 | pass→pass | 12,760 | 13,723 | +8% | 1 | 1 | 0% | 2,479 | 3,504 | +41% | 0 | 0 | — |
case-11 | pass→pass | 13,196 | 16,492 | +25% | 1 | 1 | 0% | 2,469 | 3,674 | +49% | 0 | 0 | — |
case-12 | pass→pass | 10,985 | 10,454 | -5% | 1 | 1 | 0% | 2,174 | 2,828 | +30% | 0 | 0 | — |
case-13 | pass→pass | 8,063 | 7,012 | -13% | 1 | 1 | 0% | 1,504 | 2,005 | +33% | 0 | 0 | — |
case-14 | fail→pass | 11,893 | 8,752 | -26% | 1 | 1 | 0% | 2,209 | 2,317 | +5% | 0 | 0 | — |
case-15 | pass→pass | 12,035 | 11,080 | -8% | 1 | 1 | 0% | 2,108 | 2,757 | +31% | 0 | 0 | — |
case-16 | pass→pass | 15,312 | 16,615 | +9% | 1 | 1 | 0% | 2,369 | 3,351 | +41% | 0 | 0 | — |
case-17 | pass→pass | 13,113 | 12,696 | -3% | 1 | 1 | 0% | 2,395 | 2,893 | +21% | 0 | 0 | — |
case-18 | pass→pass | 13,716 | 11,850 | -14% | 1 | 1 | 0% | 2,526 | 2,991 | +18% | 0 | 0 | — |
case-19 | pass→pass | 13,099 | 9,578 | -27% | 1 | 1 | 0% | 2,012 | 2,493 | +24% | 0 | 0 | — |
case-20 | pass→pass | 6,641 | 6,056 | -9% | 1 | 1 | 0% | 1,213 | 1,641 | +35% | 0 | 0 | — |
case-21 | pass→pass | 9,551 | 11,315 | +18% | 1 | 1 | 0% | 1,885 | 3,011 | +60% | 0 | 0 | — |
case-22 | pass→pass | 7,775 | 8,627 | +11% | 1 | 1 | 0% | 1,721 | 2,494 | +45% | 0 | 0 | — |
case-23 | pass→pass | 13,767 | 13,332 | -3% | 1 | 1 | 0% | 2,498 | 3,143 | +26% | 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 +4 percentage points is the difference between those two pass rates over the 23 comparable cases. 1 case got worse with the skill loaded, and it is 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.