When should I use the safe-over-unsafe skill?

Use it whenever you are designing types backed by raw pointers, adding unsafe blocks for performance, or implementing manual Send/Sync traits for complex types.

Can this skill help with testing unsafe code?

Yes, it provides a comprehensive verification layer guide including Miri for dynamic UB detection, Kani for formal proofs, and Proptest for logic verification.

What is the soundness boundary in Rust?

The soundness boundary is typically the module privacy limit. It ensures that even if internal code uses 'unsafe', the public API remains impossible to misuse in a way that causes undefined behavior.

How does this help with Send and Sync implementations?

It provides a specific checklist for manual Send/Sync implementations to prevent the most common real-world soundness bugs found in popular Rust crates.

Safe Rust API Designer

Name: Safe Rust API Designer
Author: ahrav

byahrav

•

セキュリティとテスト

Guides the creation of sound, safe public APIs that wrap unsafe Rust code using industry-standard memory safety and pointer handling patterns.

This skill provides a rigorous framework for Rust developers working with unsafe code, raw pointers, FFI, or manual memory management. It ensures that the 'soundness boundary' is correctly established at the module level, helping developers document safety invariants, handle variance correctly, and implement robust panic safety patterns. By following these standardized checklists and verification layers, developers can build high-performance internals without sacrificing the safety guarantees of the Rust ecosystem, effectively preventing undefined behavior and security vulnerabilities.

主な機能

01Generic parameter defense and variance (PhantomData) guidance

02Panic safety implementation patterns like decrement-before-drop

03Detailed documentation patterns for validity and safety invariants

042 GitHub stars

05Formalized safety boundary establishment via module privacy

06Multi-layered verification using Miri, Kani, and Tree Borrows

ユースケース

01Wrapping low-level C libraries (FFI) with a safe, idiomatic Rust interface

02Conducting security and soundness audits on existing unsafe Rust codebases

03Building high-performance data structures using MaybeUninit or raw pointers

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ahrav/scratch-scanner-rs safe-over-unsafe

For use in Claude.ai and ChatGPT

主な機能

01Generic parameter defense and variance (PhantomData) guidance

02Panic safety implementation patterns like decrement-before-drop

03Detailed documentation patterns for validity and safety invariants

042 GitHub stars

05Formalized safety boundary establishment via module privacy

06Multi-layered verification using Miri, Kani, and Tree Borrows

ユースケース

01Wrapping low-level C libraries (FFI) with a safe, idiomatic Rust interface

02Conducting security and soundness audits on existing unsafe Rust codebases

03Building high-performance data structures using MaybeUninit or raw pointers

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ahrav/scratch-scanner-rs safe-over-unsafe

For use in Claude.ai and ChatGPT