The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze user-supplied Rust code, error messages (e.g., E0510, E0509), and Cargo.toml metadata to provide diagnostic advice.
Ingestion points: User-provided queries containing code snippets or compiler error logs in SKILL.md (e.g., "the question or the error names").
Boundary markers: The skill does not specify delimiters (like triple backticks or unique tokens) or instructions to treat user-provided code as potentially untrusted data.
Capability inventory: The skill suggests the use of powerful command-line tools such as cargo check, cargo clippy, and cargo fix (which modifies files on disk) to verify behavior and apply migrations.
Sanitization: No explicit sanitization, validation, or escaping logic is provided for the external content before it is processed by the agent or passed to system tools.
[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute various cargo commands, including cargo fix --edition and cargo clippy. While these are standard tools for Rust development, they represent an execution capability that operates on the project environment.
Evidence: References to cargo fix --edition, cargo clippy --locked, and cargo +<msrv> check --locked throughout the documentation.

rust-pattern-semantics