メニュー
Find duplication in Rust code and consolidate it using Traits, Generics, shared functions, and macros. Use when asked to reduce Rust code duplication.
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
直接コマンドでは確認用 Prompt が省略されます。実行前にソースを確認してください。
npx skills add https://github.com/Cosmian/kms --skill rust-refactorコマンドは1行のまま表示されます。コピー前に横へスクロールして全体を確認してください。
ローカルで確認しますか？SkillsMP が現在取得できるファイルをダウンロードできます。
SKILL.md を表示中
Scan Rust code for every form of panic or brutal process exit: panic!, todo!, unimplemented!, unreachable!, .unwrap(), .expect() in production, process::exit/abort, integer overflow, and out-of-bounds indexing. Produces a ranked report with patches. Use when reviewing Rust code quality, before any PR, or when adding new Rust code.
Hardcore Rust code review gate: orchestrates ALL Rust review skills in sequence (panic audit, error propagation, async refactor, simplify, refactor, patterns, security, cryptography). Each skill writes its report to ./review/. Produces a unified go/no-go verdict. Use before submitting any significant Rust PR or after large code generation.
Write the next version entry at the top of CHANGELOG.md by summarizing all changes since the last tagged release. Use when preparing release notes.
SOC 職業分類に基づく
| name | rust-refactor |
| description | Find duplication in Rust code and consolidate it using Traits, Generics, shared functions, and macros. Use when asked to reduce Rust code duplication. |
Find opportunities to reduce LOC and consolidate duplicated behavior using Traits, Generics, shared functions, and macros.
# Find clone-heavy code
rg "\.clone\(\)" --stats | tail -5
# Spot copy-paste (identical blocks ≥ 5 lines) across target crate(s)
rg -l "fn create_" crate/server/src/core/operations/
Run the Duplication Scan checklist:
match arms that repeat the same multi-line logicimpl blocks that repeat the same methods on different typesFrom / Into candidates)if cfg!(feature = "...") guards around the same patternFor each duplicate, pick the Rust pattern:
| Smell | Pattern |
|---|---|
| Same logic on N types | Generic function / impl<T: Bound> |
| Same interface, different impls | Trait + impl Trait for Type |
Repetitive match / boilerplate impls | Declarative macro (macro_rules!) |
| Cross-cutting behaviour (logging, auth) | Blanket impl or extension trait |
| Repeated struct field group | Composition / newtype |
Repeated From/Into conversions | Derive macro or impl From<X> for Y |
| Repeated KMIP operation boilerplate | macro_rules! dispatch helpers |
Create a ranked list ordered by: impact (lines saved) ÷ risk (tests affected). Present it to the user before touching code:
[HIGH] Extract `KeyOperations` trait — saves ~120 LOC, touches 3 files
[MED] Generic `retrieve<T: KmipObject>` — saves ~60 LOC, touches 2 files
[LOW] `macro_rules! impl_kmip_response!` — saves ~30 LOC, touches 5 files
cargo clippy-all and cargo fmt --all.cargo test -p <crate>.cargo clippy-all # zero warnings
cargo fmt --all # no formatting drift
cargo test -p <crate> # affected crate(s) only
git diff --stat # every hunk explainable by the task
dyn Trait unless the set of types is open at runtime.#[allow(clippy::...)] only with an inline comment explaining why the lint cannot be satisfied.rg "fn_name" --type rust.use statements) always go at the top of the file — never inline inside function bodies.#[cfg(feature = "non-fips")] on the fn, not inside the body.// Before: duplicated retrieve logic in get.rs, export.rs, decrypt.rs
async fn get_object(uid: &str, db: &dyn Database) -> KResult<KmipObject> { ... }
async fn export_object(uid: &str, db: &dyn Database) -> KResult<KmipObject> { ... }
// After: shared generic with trait bound
async fn retrieve_object<P: Permission>(
uid: &str,
caller: &str,
db: &dyn Database,
) -> KResult<KmipObject>
where P: CheckAccess { ... }
macro_rules! dispatch_operation {
($op:expr, $kms:expr, $params:expr) => {
match $op {
Operation::Get(req) => get::get($kms, req, $params).await,
Operation::Locate(req) => locate::locate($kms, req, $params).await,
// ...
}
};
}
// Before: raw String everywhere — easy to mix up key UID and user ID
fn get_key(uid: String, owner: String) -> ...
// After: newtypes prevent accidental parameter swap
struct KeyUid(String);
struct UserId(String);
fn get_key(uid: KeyUid, owner: UserId) -> ...