skills/iliaal/ai-skills/rust-systems/Gen Agent Trust Hub

rust-systems

Pass

Audited by Gen Agent Trust Hub on Sep 21, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill documents the use of standard Rust development tools such as cargo, clippy, and cargo-nextest. It provides specific guidance on avoiding side effects during testing, such as mutating process-global state or environment variables, which prevents race conditions and accidental execution of sibling tests.
  • [INDIRECT_PROMPT_INJECTION]: The skill identifies common attack surfaces in Rust applications, such as parsing untrusted input in CLIs and HTTP services. It mandates defensive measures including schema validation via the validator crate, the use of sqlx compile-time checked macros to prevent SQL injection, and specific warnings about non-standard character handling (e.g., bare \r newlines or Unicode case folding) that could lead to security filter bypasses.
  • [DATA_EXFILTRATION]: The instructions provide patterns for secure network operations using axum and reqwest. A critical security rule is established for authentication and authorization decisions: the system must implement a "deny-by-default" fallback if an outbound security check is unavailable or returns an unrecognized response.
  • [DYNAMIC_EXECUTION]: While the skill discusses Rust macros (declarative and procedural), it focuses on hygiene and compile-time safety. It warns against using panic! or unwrap() in procedural macros, which would lead to compiler-internal failures, and instead promotes using syn::Error for better diagnostics.
  • [EXTERNAL_DOWNLOADS]: The skill mentions common, trusted CI/CD utilities such as taiki-e/install-action and Swatinem/rust-cache. These are standard tools for optimizing Rust build pipelines and are referenced in a way that aligns with industry best practices.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 21, 2026, 03:27 PM
Security Audit — agent-trust-hub — rust-systems