The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to perform mechanical quality passes using standard Rust toolchain commands, including cargo clippy --all-targets --all-features -- -D warnings, cargo fmt --all --check, and cargo test --all-targets. These are legitimate, local-only commands necessary for the primary purpose of code review. It also uses standard grep commands to audit source files for potential issues like unjustified panics or dangerous async patterns.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze external Rust source code, which constitutes an ingestion point for untrusted data. 1. Ingestion points: User-provided Rust source code files located in the src/ and tests/ directories. 2. Boundary markers: The instructions do not explicitly mandate prompt delimiters or "ignore embedded instructions" headers for the analyzed code, though it is processed by static analysis tools. 3. Capability inventory: The skill utilizes shell commands (cargo, grep) to inspect the ingested code. 4. Sanitization: The risk is mitigated through the use of standard, non-executable static analysis tools that validate the syntax and structure of the input rather than interpreting instructions embedded within it.

rust-agents-code-reviewer