The Agent Skills Directory

[SAFE]: The skill provides authoritative guidance on Rust Cargo workflows while emphasizing security best practices. It explicitly warns against hardcoding credentials, printing secrets in logs, or performing sensitive operations like publishing without authorization.
[COMMAND_EXECUTION]: The skill describes the use of standard Rust commands such as cargo run, cargo test, and cargo build. These commands are recognized as points of code execution, and the skill includes appropriate safeguards and diagnostic advice (e.g., using --locked, --frozen, or --offline flags) to manage these actions safely.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingests untrusted project data (e.g., Cargo.toml, build.rs, and source code) and possesses the capability to execute commands. However, the instructions mandate verification steps, such as using cargo metadata and cargo tree to inspect the effective state, and require explicit authorization for any external mutation, which effectively manages the risk.
Ingestion points: Processes project manifests (Cargo.toml), lockfiles, and source files across the workspace.
Boundary markers: Instructions explicitly require locating the root and confirming the effective manifest before making changes.
Capability inventory: Includes subprocess execution of the Cargo toolchain for building, testing, and running applications.
Sanitization: Relies on official Cargo validation tools and explicit human-in-the-loop authorization for sensitive tasks.

rust-cargo-build