The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands to analyze the local codebase.
Evidence: Executes git diff, grep, cargo clippy, cargo check, and cargo test to identify changed files and verify code correctness.
[REMOTE_CODE_EXECUTION]: The skill triggers the execution of local project code through standard Rust build and test tools.
Evidence: Running cargo check, cargo clippy, and cargo test (Step 3 and Post-Fix Verification) executes project-defined logic such as build.rs scripts, procedural macros, and unit tests. This represents a potential execution vector if the repository under review contains malicious code.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection from the untrusted code it processes.
Ingestion points: git diff output, Cargo.toml, and .rs source files (Steps 1, 2, and 4).
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the prompt interpolation steps.
Capability inventory: Full shell execution via cargo and git commands.
Sanitization: No sanitization or validation of the ingested code content is performed prior to processing.

review-rust