tradeoffs - SKILL.md Agent Skill

name: tradeoffs description: This skill should be used when the user wants to understand the tradeoffs between two or more Rust design approaches without being told which one to pick. Use when the user asks "what's the difference between X and Y", "should I use X or Y", "when would I choose X over Y", or is deciding between ownership strategies, async approaches, error handling patterns, or abstraction choices like trait objects vs generics. argument-hint: ""

Present the tradeoffs clearly and neutrally. The user makes the call.

For each approach (typically 2–3):

### [Approach Name]
**When it fits**: ...
**Cost / tradeoff**: ...
**Rust idiom level**: common / situational / advanced

Close with a question:

"Given [X constraint from context], which of these feels right for what you're building?"

Rc vs Arc vs Box
async fn vs sync + spawn_blocking
thiserror vs anyhow vs custom error enum
Vec<T> vs &[T] vs accepting an impl Iterator<Item=T> in function signatures
Struct with lifetimes vs owned struct vs Arc-shared fields
enum dispatch vs trait objects (dyn Trait) vs generics (impl Trait)
Mutex<T> vs message passing (mpsc) for shared mutable state
tokio::sync::Mutex vs std::sync::Mutex

Don't recommend one option unless the context makes the choice obviously clear — and if you do, explain the specific reason why
Keep each section to 3–4 lines — don't over-explain
Always highlight any Rust-specific gotcha (e.g. Rc is not Send, dyn Trait requires object safety)

$ARGUMENTS