agentleFS
kimgoetzkekimgoetzke/coding-agent-configs/skills/rust-standards/SKILL.md

rust-standards

Skill2 starsChanged 7 months agoMITView on GitHub

Write idiomatic Rust code. Use when writing Rust code.

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Show its rules and text

We split this skill into the rules it gives an agent

18 rules. About 0 are common, 6 are distinctive, and 12 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 12

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use newtypes: struct Email(String) for domain semantics

    line 20
  2. Very specific:

    Prefer slice patterns: if let [first, .., last] = slice

    line 21
  3. Very specific:

    Avoid Vec abuse: use arrays for fixed sizes

    line 23
  4. Very specific:

    Prefer &str over String in function parameters

    line 27
  5. Very specific:

    Use String for ownership: return String when transferring ownership

    line 28
  6. Very specific:

    Prefer bytes: s.bytes() over s.chars() when ASCII

    line 29
  7. Very specific:

    Use Cow<str> when you might need to modify borrowed data

    line 30
  8. Very specific:

    Use format! over string concatenation with +

    line 31
  9. Very specific:

    Avoid nested iteration: contains() on string is O(n\*m)

    line 32
  10. Very specific:

    Use ? for all fallible operations

    line 36
  11. Distinctive:

    Don't hold locks across await: use scoped guards

    line 57
  12. Distinctive:

    Avoid unless necessary: prefer functions/generics

    line 61
  13. Distinctive:

    Follow Rust syntax: macro input should look like Rust

    line 62
  14. Very specific:

    Don't restate the name: fn connect() doesn't need "Connects to the server"

    line 77
  15. Distinctive:

    Document the non-obvious: panics, errors, surprising edge cases.

    line 78
  16. Distinctive:

    No filler: no "This method...", "Note that...", "Please be aware..."

    line 79
  17. Distinctive:

    Don't write doctests purely for coverage.

    line 81
  18. Very specific:

    Use clippy with #![deny(clippy::all, clippy::pedantic)] - fix all warnings

    line 95

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Testing

  • Unit tests in #[cfg(test)] modules within each source file
  • Integration tests in tests/ directory

Naming

  • No get_ prefix: fn name() not fn get_name()
  • Iterator convention: iter() / iter_mut() / into_iter()
  • Conversion naming: as_ (cheap &), to_ (expensive), into_ (ownership)
  • Static var prefix: G_CONFIG for static, no prefix for const

Data Types

  • Use newtypes: struct Email(String) for domain semantics
  • Prefer slice patterns: if let [first, .., last] = slice
  • Pre-allocate: Vec::with_capacity(), String::with_capacity()
  • Avoid Vec abuse: use arrays for fixed sizes

Strings

  • Prefer &str over String in function parameters
  • Use String for ownership: return String when transferring ownership
  • Prefer bytes: s.bytes() over s.chars() when ASCII
  • Use Cow<str> when you might need to modify borrowed data
  • Use format! over string concatenation with +
  • Avoid nested iteration: contains() on string is O(n*m)

Error Handling

  • Use ? for all fallible operations
  • unwrap() in tests only, never production
  • expect() only for provably impossible states; the message must justify why it can't fail e.g. expect("regex is valid: validated at compile time")
  • unwrap_or / unwrap_or_else / unwrap_or_default for deliberate fallbacks
  • Assertions for invariants: assert! at function entry

Memory

  • Meaningful lifetimes: 'src, 'ctx not just 'a
  • try_borrow() for RefCell to avoid panic
  • Shadowing for transformation: let x = x.parse()?

Concurrency

  • Identify lock ordering to prevent deadlocks
  • Atomics for primitives, not Mutex for bool/usize
  • Choose memory order carefully: Relaxed/Acquire/Release/SeqCst

Async

  • Sync for CPU-bound; async is for I/O
  • Don't hold locks across await: use scoped guards

Macros

  • Avoid unless necessary: prefer functions/generics
  • Follow Rust syntax: macro input should look like Rust

Deprecated → Better

  • lazy_static! → std::sync::OnceLock (since 1.70)
  • once_cell::Lazy → std::sync::LazyLock (since 1.80)
  • std::sync::mpsc → crossbeam::channel only if you need multi-consumer or better performance under contention; std::sync::mpsc is fine for most use cases
  • failure/error-chain → thiserror/anyhow
  • try!() → ? operator (since 2018)

Docs

  • All pub methods must have /// doc comments
  • All /// doc comments must be extremely concise
  • Imperative mood: Returns the length not This function returns the length
  • Don't restate the name: fn connect() doesn't need "Connects to the server"
  • Document the non-obvious: panics, errors, surprising edge cases; skip obvious params/returns
  • No filler: no "This method...", "Note that...", "Please be aware..."
  • # Examples only for non-trivial usage; doctests must compile and pass
  • Don't write doctests purely for coverage; write them only when the example genuinely aids understanding

Quick Reference

Naming: snake_case (fn/var), CamelCase (type), SCREAMING_CASE (const)
Format: rustfmt (just use it)
Docs: /// for public items, //! for module docs
Lint: #![deny(clippy::all, clippy::pedantic)]

Miscellaneous

  • Derive Debug on all public types; derive Clone, PartialEq only when needed
  • Use clippy with #![deny(clippy::all, clippy::pedantic)] - fix all warnings
  • No unsafe blocks
  • Modules: one file per module, mod.rs only for re-exports

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.