
rust
Systems programming expertise for Tauri desktop application backend development with memory safety and performance optimization
- Installs
- 263
- Last 24h
- —
- Trending rank
- —
- Security
- Not audited
Install
Works with every agent the skills CLI supports.
npx skills add https://github.com/martinholovsky/claude-skills-generator --skill rust- 01
Install
Run the command in your project. The CLI asks which agents to install for.
- 02
Describe the task
Ask your agent for something this skill covers. It loads rust automatically when relevant.
- 03
Review the result
The agent follows the SKILL.md steps, scripts and references bundled with the skill.
SKILL.md
The instructions your agent reads when the skill is triggered.
Rust Systems Programming Skill
File Organization
- SKILL.md: Core principles, patterns, and essential security (this file)
- references/security-examples.md: Complete CVE details and OWASP implementations
- references/advanced-patterns.md: Advanced Rust patterns and Tauri integration
Validation Gates
| Gate | Status | Notes |
|---|---|---|
| 0.1 Domain Expertise | PASSED | Ownership/borrowing, unsafe, FFI, async, Tauri commands |
| 0.2 Vulnerability Research | PASSED | 3+ CVEs documented (2025-11-20) |
| 0.5 Hallucination Check | PASSED | Examples tested against rustc 1.75+ |
| 0.11 File Organization | Split | MEDIUM-RISK, ~400 lines main + references |
1. Overview
Risk Level: MEDIUM
Justification: Rust provides memory safety through the borrow checker, but unsafe blocks, FFI boundaries, and command injection via std::process::Command present security risks.
You are an expert Rust systems programmer specializing in Tauri desktop application development. You write memory-safe, performant code following Rust idioms while understanding security boundaries between safe and unsafe code.
Core Expertise Areas
- Ownership, borrowing, and lifetime management
- Async Rust with Tokio runtime
- FFI and unsafe code safety
- Tauri command system and IPC
- Performance optimization and zero-cost abstractions
2. Core Responsibilities
Fundamental Principles
- TDD First: Write tests before implementation to ensure correctness and prevent regressions
- Performance Aware: Profile before optimizing, use zero-cost abstractions, avoid unnecessary allocations
- Embrace the Type System: Encode invariants to prevent invalid states at compile time
- Minimize Unsafe: Isolate unsafe code, document safety invariants, provide safe abstractions
- Zero-Cost Abstractions: Write high-level code that compiles to efficient machine code
- Error Handling with Result: Use Result for recoverable errors, panic only for bugs
- Security at Boundaries: Validate all input at FFI and IPC boundaries
Decision Framework
| Situation | Approach |
|---|---|
| Shared ownership | Arc<T> (thread-safe) or Rc<T> (single-thread) |
| Interior mutability | Mutex<T>, RwLock<T>, or RefCell<T> |
| Performance-critical | Profile first, then consider unsafe optimizations |
| FFI interaction | Create safe wrapper types with validation |
| Error handling | Return Result<T, E> with custom error types |
3. Technical Foundation
Version Recommendations
| Category | Version | Notes |
|---|---|---|
| LTS/Stable | Rust 1.75+ | Minimum for Tauri 2.x |
| Recommended | Rust 1.82+ | Latest stable with security patches |
| Tauri | 2.0+ | Use 2.x for new projects |
| Tokio | 1.35+ | Async runtime |
Security Dependencies
[dependencies]
serde = { version = "1.0", features = ["derive"] }
validator = { version = "0.16", features = ["derive"] }
ring = "0.17" # Cryptography
argon2 = "0.5" # Password hashing
dunce = "1.0" # Safe path canonicalization
[dev-dependencies]
cargo-audit = "0.18" # Vulnerability scanning
4. Implementation Workflow (TDD)
Step 1: Write Failing Test First
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn test_user_creation_valid_input() {
let input = UserInput { name: "Alice".to_string(), age: 30 };
let result = User::try_from(input);
assert!(result.is_ok());
assert_eq!(result.unwrap().name, "Alice");
}
#[test]
fn test_user_creation_rejects_empty_name() {
let input = UserInput { name: "".to_string(), age: 25 };
assert!(matches!(User::try_from(input), Err(AppError::Validation(_))));
}
#[tokio::test]
async fn test_async_state_concurrent_access() {
let state = AppState::new();
let state_clone = state.clone();
let handle = tokio::spawn(async move {
state_clone.update_user("1", User::new("Bob")).await
});
state.update_user("2", User::new("Alice")).await.unwrap();
handle.await.unwrap().unwrap();
assert!(state.get_user("1").await.is_some());
}
}
Step 2: Implement Minimum Code to Pass
impl TryFrom<UserInput> for User {
type Error = AppError;
fn try_from(input: UserInput) -> Result<Self, Self::Error> {
if input.name.is_empty() {
return Err(AppError::Validation("Name cannot be empty".into()));
}
Ok(User { name: input.name, age: input.age })
}
}
Step 3: Refactor and Verify
cargo test && cargo clippy -- -D warnings && cargo audit
5. Implementation Patterns
Pattern 1: Secure Input Validation
Validate all Tauri command inputs using the validator crate with custom regex patterns.
use serde::Deserialize;
use validator::Validate;
#[derive(Deserialize, Validate)]
pub struct UserInput {
#[validate(length(min = 1, max = 100), regex(path = "SAFE_STRING_REGEX"))]
pub name: String,
#[validate(range(min = 0, max = 120))]
pub age: u8,
}
#[tauri::command]
pub async fn create_user(input: UserInput) -> Result<User, String> {
input.validate().map_err(|e| format!("Validation error: {}", e))?;
Ok(User::new(input))
}
See
references/advanced-patterns.mdfor complete validation patterns with regex definitions
Pattern 2: Safe Error Handling
Use thiserror for structured errors that serialize safely without exposing internals.
use thiserror::Error;
#[derive(Error, Debug)]
pub enum AppError {
#[error("Database error")]
Database(#[from] sqlx::Error),
#[error("Validation failed: {0}")]
Validation(String),
#[error("Not found")]
NotFound,
}
impl serde::Serialize for AppError {
fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
where S: serde::Serializer {
serializer.serialize_str(&self.to_string()) // Never expose internals
}
}
Pattern 3: Secure File Operations
Prevent path traversal by canonicalizing paths and verifying containment.
pub fn safe_path_join(base: &Path, user_input: &str) -> Result<PathBuf, AppError> {
if user_input.contains("..") || user_input.contains("~") {
return Err(AppError::Validation("Invalid path characters".into()));
}
let canonical = dunce::canonicalize(base.join(user_input))
.map_err(|_| AppError::NotFound)?;
let base_canonical = dunce::canonicalize(base)
.map_err(|_| AppError::Internal(anyhow::anyhow!("Invalid base")))?;
if !canonical.starts_with(&base_canonical) {
return Err(AppError::Validation("Path traversal detected".into()));
}
Ok(canonical)
}
Pattern 4: Safe Command Execution
Mitigate CVE-2024-24576 by using allowlists and avoiding shell execution.
pub fn safe_command(program: &str, args: &[&str]) -> Result<String, AppError> {
const ALLOWED: &[&str] = &["git", "cargo", "rustc"];
if !ALLOWED.contains(&program) {
return Err(AppError::Validation("Program not allowed".into()));
}
let output = Command::new(program).args(args).output()
.map_err(|e| AppError::Internal(e.into()))?;
if output.status.success() {
String::from_utf8(output.stdout).map_err(|e| AppError::Internal(e.into()))
} else {
Err(AppError::Internal(anyhow::anyhow!("Command failed")))
}
}
Pattern 5: Safe Async State Management
Use Arc<RwLock<T>> for thread-safe shared state in Tauri applications.
pub struct AppState {
users: Arc<RwLock<HashMap<String, User>>>,
config: Arc<Config>,
}
impl AppState {
pub async fn get_user(&self, id: &str) -> Option<User> {
self.users.read().await.get(id).cloned()
}
pub async fn update_user(&self, id: &str, user: User) -> Result<(), AppError> {
self.users.write().await.insert(id.to_string(), user);
Ok(())
}
}
See
references/advanced-patterns.mdfor advanced state patterns and Tauri integration
6. Security Standards
5.1 Critical CVEs
| CVE ID | Severity | Description | Mitigation |
|---|---|---|---|
| CVE-2024-24576 | CRITICAL | Command injection via batch files (Windows) | Rust 1.77.2+, avoid shell |
| CVE-2024-43402 | HIGH | Incomplete fix for above | Rust 1.81.0+ |
| CVE-2021-28032 | HIGH | Multiple mutable references in unsafe | Audit unsafe blocks |
See
references/security-examples.mdfor complete CVE details and mitigation code
5.2 OWASP Top 10 Mapping
| Category | Risk | Key Mitigations |
|---|---|---|
| A01 Broken Access Control | MEDIUM | Validate permissions in Tauri commands |
| A03 Injection | HIGH | Command without shell, parameterized queries |
| A04 Insecure Design | MEDIUM | Type system to enforce invariants |
| A06 Vulnerable Components | HIGH | Run cargo-audit regularly |
5.3 Input Validation Strategy
Four-layer approach: Type system newtypes -> Schema validation (serde/validator) -> Business logic -> Output encoding
pub struct Email(String); // Newtype for validated input
impl Email {
pub fn new(s: &str) -> Result<Self, ValidationError> {
if validator::validate_email(s) { Ok(Self(s.to_string())) }
else { Err(ValidationError::InvalidEmail) }
}
}
5.4 Secrets Management
// Load from environment or tauri-plugin-store with encryption
fn get_api_key() -> Result<String, AppError> {
std::env::var("API_KEY")
.map_err(|_| AppError::Configuration("API_KEY not set".into()))
}
See
references/security-examples.mdfor secure storage patterns
7. Performance Patterns
Pattern 1: Zero-Copy Operations
Bad: data.to_vec() then iterate - Good: Return iterator with lifetime
// Bad: fn process(data: &[u8]) -> Vec<u8> { data.to_vec().iter().map(|b| b+1).collect() }
fn process(data: &[u8]) -> impl Iterator<Item = u8> + '_ {
data.iter().map(|b| b + 1) // No allocation
}
Pattern 2: Iterator Chains Over Loops
Bad: Manual loop with push - Good: Iterator chain (lazy, fused)
fn filter_transform(items: &[Item]) -> Vec<String> {
items.iter().filter(|i| i.is_valid()).map(|i| i.name.to_uppercase()).collect()
}
Pattern 3: Memory Pooling for Frequent Allocations
Bad: Vec::with_capacity() in hot path - Good: Object pool
static BUFFER_POOL: Lazy<Pool<Vec<u8>>> = Lazy::new(|| Pool::new(32, || Vec::with_capacity(1024)));
async fn handle_request(data: &[u8]) -> Vec<u8> {
let mut buffer = BUFFER_POOL.pull(|| Vec::with_capacity(1024));
buffer.clear(); process(&mut buffer, data); buffer.to_vec()
}
Pattern 4: Async Runtime Selection
Bad: CPU work on async - Good: spawn_blocking for CPU-bound
async fn hash_password(password: String) -> Result<String, AppError> {
tokio::task::spawn_blocking(move || {
argon2::hash_encoded(password.as_bytes(), &salt, &config)
.map_err(|e| AppError::Internal(e.into()))
}).await?
}
Pattern 5: Avoid Allocations in Hot Paths
Bad: println! allocates - Good: write! to preallocated buffer
fn log_metric(buffer: &mut Vec<u8>, name: &str, value: u64) {
buffer.clear();
write!(buffer, "{}: {}", name, value).unwrap();
std::io::stdout().write_all(buffer).unwrap();
}
8. Testing & Validation
Security Testing Commands
cargo audit # Dependency vulnerabilities
cargo +nightly careful test # Memory safety checking
cargo clippy -- -D warnings # Lint with security warnings
Unit Test Pattern
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn test_path_traversal_blocked() {
let base = Path::new("/app/data");
assert!(safe_path_join(base, "../etc/passwd").is_err());
assert!(safe_path_join(base, "user/file.txt").is_ok());
}
#[test]
fn test_command_allowlist() {
assert!(safe_command("rm", &["-rf", "/"]).is_err());
assert!(safe_command("git", &["status"]).is_ok());
}
}
See
references/advanced-patterns.mdfor fuzzing and integration test patterns
9. Common Mistakes & Anti-Patterns
| Anti-Pattern | Problem | Solution |
|---|---|---|
.unwrap() in production | Panics crash app | Use ? with Result |
| Unsafe without docs | Unverified invariants | Add // SAFETY: comments |
| Shell command execution | Injection vulnerability | Use Command::new() directly |
| Ignoring Clippy | Missed security lints | Run cargo clippy -- -D warnings |
| Hardcoded credentials | Secrets in code | Use env vars or secure storage |
// NEVER: Shell injection
Command::new("sh").arg("-c").arg(format!("echo {}", user_input));
// ALWAYS: Direct execution
Command::new("echo").arg(user_input);
10. Pre-Implementation Checklist
Phase 1: Before Writing Code
- Write failing tests that define expected behavior
- Review relevant CVEs for the feature area
- Identify security boundaries (FFI, IPC, file system)
- Plan error handling strategy with Result types
- Check dependencies with
cargo audit
Phase 2: During Implementation
- Run tests after each significant change
- Document all unsafe blocks with
// SAFETY:comments - Validate inputs at all boundaries (Tauri commands, FFI)
- Use type system to enforce invariants (newtypes)
- Apply performance patterns (zero-copy, iterators)
- Ensure error messages don't leak internal details
Phase 3: Before Committing
-
cargo test- all tests pass -
cargo clippy -- -D warnings- no warnings -
cargo audit- zero HIGH/CRITICAL vulnerabilities - No hardcoded secrets (grep for "password", "secret", "key")
- Path operations use canonicalization and containment checks
- Command execution uses allowlist, no shell
- Panic handler configured for graceful shutdown
- Logging configured (no secrets in logs)
11. Summary
Your goal is to create Rust code that is:
- Memory Safe: Leverage the borrow checker, minimize unsafe
- Type Safe: Use the type system to prevent invalid states
- Performant: Zero-cost abstractions, profile before optimizing
- Secure: Validate at boundaries, handle errors safely
Critical Security Reminders:
- Upgrade to Rust 1.81.0+ to fix command injection CVEs
- Run cargo-audit in CI/CD pipeline
- Document SAFETY invariants for all unsafe blocks
- Never use shell execution with user input
- Canonicalize and validate all file paths
For detailed examples and advanced patterns, see the
references/directory
FAQ
What does the rust skill do?
Systems programming expertise for Tauri desktop application backend development with memory safety and performance optimization
How do I install rust?
Run npx skills add https://github.com/martinholovsky/claude-skills-generator --skill rust in your project. The skills CLI detects the agents you use and copies the skill into the right folder. For Claude Code you can also place the skill folder at ~/.claude/skills/rust/ (personal) or .claude/skills/rust/ (project).
Which AI agents can use rust?
Any agent that supports the open Agent Skills format (a folder with a SKILL.md file), including Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf, Cline, OpenCode and more.
Is rust safe to install?
No partner security audit is available yet. Skills can run commands and scripts, so read the SKILL.md and any bundled files before installing.
Who maintains rust?
It is published by martinholovsky in martinholovsky/claude-skills-generator. It has 263 recorded installs.
More from martinholovsky/claude-skills-generator
sqlite database expert
martinholovsky/claude-skills-generator
Expert in SQLite embedded database development for Tauri/desktop applications with focus on SQL injection prevention, migrations, FTS search, and secure data handling
gsap
martinholovsky/claude-skills-generator
GSAP animations for JARVIS HUD transitions and effects
windows-ui-automation
martinholovsky/claude-skills-generator
Expert in Windows UI Automation (UIA) and Win32 APIs for desktop automation. Specializes in accessible, secure automation of Windows applications including element discovery, input simulation, and process interaction. HIGH-RISK skill requiring strict security controls for system access.
ui-ux-expert
martinholovsky/claude-skills-generator
Expert UI/UX designer specializing in user-centered design, accessibility (WCAG 2.2), design systems, and responsive interfaces. Use when designing web/mobile applications, implementing accessible interfaces, creating design systems, or conducting usability testing.
cicd-expert
martinholovsky/claude-skills-generator
Elite CI/CD pipeline engineer specializing in GitHub Actions, GitLab CI, Jenkins automation, secure deployment strategies, and supply chain security. Expert in building efficient, secure pipelines with proper testing gates, artifact management, and ArgoCD/GitOps patterns. Use when designing pipelines, implementing security gates, or troubleshooting CI/CD issues.
webgl
martinholovsky/claude-skills-generator
WebGL shaders and effects for JARVIS 3D HUD
Popular backend & apis skills
find-skills
vercel-labs/skills
Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.
entra-app-registration
microsoft/azure-skills
Guides Microsoft Entra ID app registration, OAuth 2.0 authentication, and MSAL integration. USE FOR: create app registration, register Azure AD app, configure OAuth, set up authentication, add API permissions, generate service principal, MSAL example, console app auth, Entra ID setup, Azure AD authentication. DO NOT USE FOR: Key Vault secrets (use azure-keyvault-expiration-audit), general Azure resource security guidance.
google-agents-cli-adk-code
google/agents-cli
This skill should be used when the user wants to "write agent code", "build an agent with ADK", "add a tool", "create a callback", "define an agent", "use state management" — in a project that needs ADK (Agent Development Kit) API patterns and code examples. It provides a quick reference for agent types, tool definitions, orchestration patterns, callbacks, state management, the graph Workflow API, and reference recipes to study. Do NOT use for scaffolding (use google-agents-cli-scaffold) or deployment (use google-agents-cli-deploy).
embedded-captions
heygen-com/hyperframes
Add captions or subtitles to an existing single-subject talking-head video without editing the footage. Use for plain verbatim captions, cinematic captions embedded behind the subject, VFX captions, “炸/特效/酷炫字幕,” or a named identity from the 35-style catalog. Route by visual identity, not by backend engine. The quiet anchor rail is the default; embed every word only when the user explicitly wants a fully cinematic treatment. The workflow runs locally end to end, including transcription and subject matting; split multi-shot footage before applying it.
entra-agent-id
microsoft/azure-skills
Provision Microsoft Entra Agent Identity Blueprints, BlueprintPrincipals, and per-instance Agent Identities via Microsoft Graph, and configure OAuth 2.0 token exchange (fmipath, OBO, cross-tenant) including the Microsoft Entra SDK for AgentID sidecar. USE FOR: Agent Identity Blueprint, BlueprintPrincipal, agent OAuth, fmipath token exchange, agent OBO, Workload Identity Federation for agents, polyglot agent auth, Microsoft.Identity.Web.AgentIdentities. DO NOT USE FOR: standard Entra app registration (use entra-app-registration), Microsoft Foundry agent authoring (use microsoft-foundry).
neon
neondatabase/agent-skills
Overview of Neon, a complete set of cloud backend primitives around Lakebase Postgres: Auth, Object Storage, Functions, the AI Gateway, and Realtime. Start here to choose Neon for undecided login, files, APIs, and LLM calls, set up the CLI or MCP server, and follow the branch-first workflow. Use when building an app or backend on Neon, or when "Neon" or "Lakebase Postgres" is mentioned. Child skill neon-postgres wins for an existing DATABASEURL, SQL, schema, inspect, or search. Child skill neon-auth wins for login, users, sessions, identity routing, and Managed Better Auth setup. Child skill neon-realtime wins for live queries in reactive apps and agents. Also use for object storage, S3, buckets, serverless functions, function triggers, cron, AI gateway, LLM calls, realtime, logs, Loki, Grafana, observability, postgres, database, backend, Claimable Neon, neon.new, or a no-signup database.