agentleFS
uswuthuswuth/rustoria/SKILL.md

rustoria

Skill0 starsChanged 35 days agoMITView on GitHub

Production-grade Rust engineering guidance for AI agents. Use for designing, implementing, reviewing, or debugging Rust code.

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 0 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

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

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

    Before recommending any dependency, check whether std already provides the capability — see rust-std and the native-first rule in …

    line 20
  2. Very specific:

    rayon is for CPU-bound work: Call it from spawn_blocking, never directly inside an async fn.

    line 70
  3. Very specific:

    Native first: Prefer std when it provides the capability.

    line 103
  4. Very specific:

    No blocking in async: Use spawn_blocking or async-native crates.

    line 104

≈ “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

rustoria

Router and entry point for the rustoria skill system. Routes to the right skill, then to the right reference.

How to use

  1. Load this router for any Rust task.
  2. Route to the skill matching the task (catalog below).
  3. Skills point to references/ for deep dives; load the reference when the task needs it.
  4. For library work, load the library skill and the native skills it builds on (e.g. axum builds on tokio and serde).

Native Rust first

Rust language, standard library, and Cargo/toolchain knowledge is the foundation. Third-party crates are extensions. Before recommending any dependency, check whether std already provides the capability — see rust-std and the native-first rule in rust-production.

Skill catalog (14 skills)

Native core

SkillUse when
rust-fundamentalsLearning or reviewing Rust language basics: ownership, borrowing, lifetimes, types, traits, generics, macros.
rust-stdChoosing or using standard-library APIs: collections, iterators, strings, fs/io, process, net, sync, conversion traits.
rust-productionDesigning, reviewing, or shipping production Rust: architecture, errors, concurrency, performance, testing, observability.
rust-debuggingDebugging Rust code, reading compiler errors, or investigating runtime failures.
rust-unsafe-checkerWriting or reviewing unsafe Rust, FFI, or raw-pointer code.
rust-dependenciesManaging dependencies, versions, SemVer, Cargo.lock, MSRV, and feature flags.

Security

SkillUse when
rust-code-auditAuditing a repository for security vulnerabilities, secret exposure, or supply-chain risks.

Library extensions

SkillUse when
rust-tokioAsync runtimes, tasks, channels, timers, cancellation.
rust-axumBuilding HTTP APIs and services with axum.
rust-serdeSerialization and deserialization with serde.
rust-dieselDatabases with Diesel ORM (schema, queries, migrations, pooling).
rust-rayonData parallelism and CPU-bound work.
rust-clapCommand-line interfaces and argument parsing.
rust-ratatuiTerminal user interfaces (TUI).

Ecosystem orientation

LayerCrateSkill
Async runtimetokio 1.xrust-tokio
HTTPaxum 0.8rust-axum
Serializationserde 1.xrust-serde
Databasediesel 2.3rust-diesel
Parallelismrayon 1.12rust-rayon
CLIclap 4.xrust-clap
TUIratatui 0.30rust-ratatui

How the layers connect:

  • axum runs on tokio. axum::serve takes a tokio::net::TcpListener; #[tokio::main] sets up the runtime.
  • diesel is synchronous. In async servers, use a connection pool (r2d2) with spawn_blocking, or diesel-async.
  • rayon is for CPU-bound work. Call it from spawn_blocking, never directly inside an async fn.
  • serde is the data layer. axum's Json extractor/response is serde; diesel models derive serde traits when needed.

References

Deep-dive documents, reachable from the skills that need them:

  • references/ownership.md — moves, Copy vs Clone, drop order, Box/Rc/Arc, Weak, leaks
  • references/borrowing.md — borrow rules, lifetimes, elision, RefCell/Cell/OnceLock/LazyLock
  • references/type-design.md — newtypes, enums, typestate, PhantomData, Deref guidance
  • references/error-handling.md — Result/Option, thiserror vs anyhow, ?, error design
  • references/concurrency.md — threads, channels, Mutex/RwLock, atomics, Send/Sync, poisoning
  • references/async.md — Future/Pin, tokio runtime, spawn, cancellation, backpressure
  • references/performance.md — allocations, clones, iterators, profiling, release profiles
  • references/testing.md — unit/integration/doc tests, property tests, async tests, coverage
  • references/security.md — input validation, secrets, injection, overflow, unsafe policy
  • references/production.md — observability, config, graceful shutdown, deployment
  • references/architecture.md — module boundaries, workspaces, feature-oriented structure
  • references/toolchain.md — cargo, rustfmt, clippy, rustdoc, features, editions, MSRV

Examples

Runnable, tested projects under examples/:

  • std-file-processor — zero-dependency std-only tool (fs, io, iterators, custom errors)
  • minimal-api — axum HTTP API
  • modular-monolith — feature-oriented module boundaries
  • async-service — tokio workers, scheduler, graceful shutdown
  • cli — clap subcommands with tests
  • library — validated newtypes with thiserror

House rules

  • Native first. Prefer std when it provides the capability; add dependencies intentionally.
  • No blocking in async. Use spawn_blocking or async-native crates.
  • No casual unwrap()/expect() in production paths. Examples may use them only where clearly demo-grade.
  • Version-sensitive advice names its version. Skills state which Rust version or crate version a recommendation applies to.
  • Editions. Examples use edition 2021. Skills note when guidance is edition-specific.

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.