MCPInstall

rust-unsafe

Rust unsafe code skill for systems programming. Use when writing or reviewing unsafe Rust, understanding what operations require unsafe, implementing safe abstractions over unsafe code, auditing unsafe blocks, or understanding raw pointers, transmute, and extern. Activates on queries about unsafe Rust, raw pointers, transmute, unsafe blocks, writing safe wrappers, UnsafeCell, unsafe trait impl, or auditing unsafe code.

Installs
560
Last 24h
—
Trending rank
—
Security
Not audited

Install

Works with every agent the skills CLI supports.

npx skills add https://github.com/mohitmishra786/low-level-dev-skills --skill rust-unsafe
  1. 01

    Install

    Run the command in your project. The CLI asks which agents to install for.

  2. 02

    Describe the task

    Ask your agent for something this skill covers. It loads rust-unsafe automatically when relevant.

  3. 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 unsafe

Purpose

Guide agents through writing, reviewing, and reasoning about unsafe Rust: what operations require unsafe, how to write safe abstractions, audit patterns, common pitfalls, and when to reach for unsafe.

Triggers

  • "When do I need to use unsafe in Rust?"
  • "How do I write a safe abstraction over unsafe code?"
  • "How do I audit an unsafe block?"
  • "What are the rules for raw pointers in Rust?"
  • "What does transmute do and when is it safe?"
  • "How do I implement UnsafeCell correctly?"

Workflow

1. The five unsafe superpowers

unsafe grants exactly five capabilities not available in safe Rust:

  1. Dereference raw pointers (*const T, *mut T)
  2. Call unsafe functions (including extern "C" functions)
  3. Access or modify mutable static variables
  4. Implement unsafe traits (Send, Sync)
  5. Access fields of unions

Everything else in Rust — including memory allocation, borrowing, closures — follows safe rules even inside unsafe blocks.

2. Raw pointers

// Creating raw pointers (safe — no dereference yet)
let x = 42u32;
let ptr: *const u32 = &x;
let mut_ptr: *mut u32 = &mut some_val as *mut u32;

// Null pointer
let null: *const u32 = std::ptr::null();
let null_mut: *mut u32 = std::ptr::null_mut();

// Dereference (unsafe)
let val = unsafe { *ptr };

// Null check
if !ptr.is_null() {
    let val = unsafe { *ptr };
}

// Offset (safe to compute, unsafe to dereference)
let arr = [1u32, 2, 3, 4, 5];
let p = arr.as_ptr();
let third = unsafe { *p.add(2) };   // arr[2]
let also_third = unsafe { *p.offset(2) };

// Slice from raw parts
let slice: &[u32] = unsafe {
    std::slice::from_raw_parts(p, arr.len())
};

Rules for sound raw pointer dereference:

  • Pointer must be non-null
  • Pointer must be aligned for T
  • Memory must be initialized for T
  • Must not violate aliasing rules (only one &mut to a location)
  • Memory must be valid for the lifetime of the reference

3. unsafe functions and traits

// Declare unsafe function (callers must uphold invariants)
/// # Safety
/// `ptr` must be non-null and aligned to `T`, and point to initialized data.
/// The caller must ensure no other mutable reference to the same location exists.
unsafe fn read_ptr<T>(ptr: *const T) -> T {
    ptr.read()  // ptr::read is unsafe
}

// Call unsafe function
let val = unsafe { read_ptr(some_ptr) };

// Unsafe trait — implementor must uphold safety invariants
unsafe trait MyUnsafeTrait {
    fn operation(&self);
}

// Implementing an unsafe trait is unsafe
unsafe impl MyUnsafeTrait for MyType {
    fn operation(&self) { /* must uphold the trait's invariants */ }
}

// Send and Sync
// Send: type can be moved to another thread
// Sync: type can be shared between threads (&T is Send)
unsafe impl Send for MyType {}
unsafe impl Sync for MyType {}

4. Safe abstractions over unsafe

// The golden rule: unsafe blocks should be small, isolated, and
// wrapped in a safe API that maintains the invariant

pub struct MyVec<T> {
    ptr: *mut T,
    len: usize,
    cap: usize,
}

impl<T> MyVec<T> {
    pub fn new() -> Self {
        MyVec { ptr: std::ptr::NonNull::dangling().as_ptr(), len: 0, cap: 0 }
    }

    // Safe public API
    pub fn get(&self, index: usize) -> Option<&T> {
        if index < self.len {
            // Safety: index < len guarantees ptr+index is in bounds and initialized
            Some(unsafe { &*self.ptr.add(index) })
        } else {
            None
        }
    }

    // # Safety comment documents the invariant
    pub fn push(&mut self, val: T) {
        if self.len == self.cap {
            self.grow();
        }
        // Safety: len < cap after grow(), so ptr+len is in bounds
        unsafe { self.ptr.add(self.len).write(val) };
        self.len += 1;
    }
}

// Implement Drop to clean up
impl<T> Drop for MyVec<T> {
    fn drop(&mut self) {
        // Safety: ptr was allocated with this layout, and all elements are initialized
        unsafe {
            std::ptr::drop_in_place(std::slice::from_raw_parts_mut(self.ptr, self.len));
            std::alloc::dealloc(self.ptr as *mut u8,
                std::alloc::Layout::array::<T>(self.cap).unwrap());
        }
    }
}

5. transmute

// transmute: reinterpret bits of one type as another
// Both types must have the same size

// Safe uses:
let x: u32 = 0x3f800000;
let f: f32 = unsafe { std::mem::transmute(x) };  // bits → float

// Transmute slice pointer (sound if types have same size/align)
let bytes: &[u8] = &[0x00, 0x00, 0x80, 0x3f];
let floats: &[f32] = unsafe {
    std::slice::from_raw_parts(bytes.as_ptr() as *const f32, 1)
};

// Prefer safe alternatives when available:
let f = f32::from_bits(x);         // instead of transmute for float bits
let n = u32::from_ne_bytes(bytes); // instead of transmute for byte arrays

Common transmute pitfalls:

  • Wrong sizes (compile error, but check for generic types)
  • Creating invalid enum values
  • Creating references with wrong lifetimes

6. UnsafeCell — interior mutability

use std::cell::UnsafeCell;

// UnsafeCell is the only way to mutate through a shared reference
struct MyCell<T> {
    value: UnsafeCell<T>,
}

impl<T: Copy> MyCell<T> {
    fn new(val: T) -> Self {
        MyCell { value: UnsafeCell::new(val) }
    }

    fn get(&self) -> T {
        // Safety: single-threaded, no concurrent mutation
        unsafe { *self.value.get() }
    }

    fn set(&self, val: T) {
        // Safety: single-threaded, no outstanding references
        unsafe { *self.value.get() = val }
    }
}

7. Unsafe audit checklist

When reviewing an unsafe block:

  • Is there a // Safety: comment explaining the invariant?
  • Is the raw pointer non-null?
  • Is the raw pointer correctly aligned for the target type?
  • Is the memory initialized?
  • Is the lifetime of the reference valid?
  • Are aliasing rules respected (no simultaneous & and &mut)?
  • For extern "C": are C invariants documented and verified?
  • For Send/Sync impl: is thread safety actually guaranteed?
  • Is the unsafe block as small as possible?
  • Is there a test under Miri for the unsafe code?

8. When to use unsafe

Before reaching for unsafe, check:
├── Does std have a safe API? (Vec, Box, Arc — usually yes)
├── Does a crate handle it? (memmap2, nix, windows-sys)
├── Can you restructure to avoid it?
└── Is the performance gain measured and significant?

Legitimate uses:
├── FFI to C libraries (extern "C")
├── OS-level APIs (syscalls, mmap, ioctl)
├── Performance-critical data structures (custom allocators, SoA)
├── Hardware access (embedded, drivers)
└── Implementing safe abstractions (the standard library itself)

For unsafe patterns and audit examples, see references/unsafe-patterns.md.

Related skills

  • Use skills/rust/rust-sanitizers-miri — Miri is the essential tool for testing unsafe code
  • Use skills/rust/rust-ffi for unsafe patterns in FFI contexts
  • Use skills/rust/rust-debugging for debugging panics in unsafe code
  • Use skills/low-level-programming/memory-model for aliasing and memory ordering in unsafe

FAQ

What does the rust-unsafe skill do?

Rust unsafe code skill for systems programming. Use when writing or reviewing unsafe Rust, understanding what operations require unsafe, implementing safe abstractions over unsafe code, auditing unsafe blocks, or understanding raw pointers, transmute, and extern. Activates on queries about unsafe Rust, raw pointers, transmute, unsafe blocks, writing safe wrappers, UnsafeCell, unsafe trait impl, or auditing unsafe code.

How do I install rust-unsafe?

Run npx skills add https://github.com/mohitmishra786/low-level-dev-skills --skill rust-unsafe 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-unsafe/ (personal) or .claude/skills/rust-unsafe/ (project).

Which AI agents can use rust-unsafe?

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-unsafe 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-unsafe?

It is published by mohitmishra786 in mohitmishra786/low-level-dev-skills. It has 560 recorded installs.

More from mohitmishra786/low-level-dev-skills

View all →

cmake

mohitmishra786/low-level-dev-skills

CMake build system skill for C/C++ projects. Use when writing or refactoring CMakeLists.txt, configuring out-of-source builds, selecting generators (Ninja, Make, VS), managing targets and dependencies with targetlinklibraries, integrating external packages via findpackage or FetchContent, enabling sanitizers, setting up toolchain files for cross-compilation, or exporting CMake packages. Activates on queries about CMakeLists.txt, cmake configure errors, target properties, install rules, CPack, or CMake presets.

static-analysis

mohitmishra786/low-level-dev-skills

Static analysis skill for C/C++ codebases. Use when hardening code quality, triaging noisy builds, running clang-tidy, cppcheck, or scan-build, interpreting check categories, suppressing false positives, or integrating static analysis into CI. Activates on queries about clang-tidy checks, cppcheck, scan-build, compilecommands.json, code hardening, or static analysis warnings.

DevOps & Cloud1.2K installs

gdb

mohitmishra786/low-level-dev-skills

GDB debugger skill for C/C++ programs. Use when starting a GDB session, setting breakpoints, stepping through code, inspecting variables, debugging crashes, using reverse debugging (record/replay), remote debugging with gdbserver, or loading core dumps. Activates on queries about GDB commands, segfaults, hangs, watchpoints, conditional breakpoints, pretty-printers, Python GDB scripting, or multi-threaded debugging.

llvm

mohitmishra786/low-level-dev-skills

LLVM IR and pass pipeline skill. Use when working directly with LLVM Intermediate Representation (IR), running opt passes, generating IR with llc, inspecting or writing LLVM IR for custom passes, or understanding how the LLVM backend lowers IR to assembly. Activates on queries about LLVM IR, opt, llc, llvm-dis, LLVM passes, IR transformations, or building LLVM-based tools.

Backend & APIs901 installs

sanitizers

mohitmishra786/low-level-dev-skills

Compiler sanitizer skill for runtime bug detection in C/C++. Use when enabling and interpreting AddressSanitizer (ASan), UndefinedBehaviorSanitizer (UBSan), ThreadSanitizer (TSan), MemorySanitizer (MSan), or LeakSanitizer (LSan) with GCC or Clang. Activates on queries about sanitizer flags, sanitizer reports, ASANOPTIONS, memory errors, data races, undefined behaviour, uninitialised reads, or choosing which sanitizer to use for a given bug class.

core-dumps

mohitmishra786/low-level-dev-skills

Core dump analysis skill for production crash triage. Use when loading core files in GDB or LLDB, enabling core dump generation on Linux/macOS, mapping symbols with debuginfo or debuginfod, or extracting backtraces from crashes without re-running the program. Activates on queries about core files, ulimit, coredumpctl, debuginfod, crash triage, or analyzing segfaults from production binaries.

DevOps & Cloud767 installs

Popular backend & apis skills

View category →

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.

OfficialBackend & APIs4.1M installs

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.

OfficialBackend & APIs566.7K installs

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).

Backend & APIs514.8K installs

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.

Backend & APIs453.3K installs

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).

OfficialBackend & APIs336.6K installs

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.

Backend & APIs237.1K installs