unsafe-rust-review

The skill consists of a set of rigorous guidelines and test cases for reviewing unsafe Rust code. It emphasizes formal logic, evidence classification, and adherence to authoritative sources like the Rust Reference.\n\n- Safety Auditing Framework: The instructions define a strict methodology for evaluating safety documentation and comments as formal proofs. This promotes high standards for code review and maintains project invariants.\n\n- Trusted External References: The skill links to official Rust documentation, the Rustonomicon, and established Google-maintained audit repositories. These references are from trusted organizations and well-known services, facilitating access to authoritative safety information.\n\n- Adversarial Modeling of Rust Code: The skill correctly identifies that safe caller-provided code should be treated as potentially adversarial in a memory-safety context, which is a best practice in secure Rust development.\n\n- Intentional Soundness Testcases: Included Rust files contain intentional safety violations. These are clearly defined within an evaluation suite to test the agent's auditing capabilities and do not pose a risk to the execution environment.