The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes instructions to run local Python scripts (audit_cargo_perf.py, check_unsafe_justifications.py, scan_rust_hotpaths.py) which analyze a repository path provided by the user. These scripts are limited to reading local files and printing findings to the console; they do not perform network operations or modify the system.
[DYNAMIC_EXECUTION]: The script scripts/audit_cargo_perf.py uses importlib.import_module("tomllib") to load the Python standard library's TOML parser. This is a standard implementation for feature detection in Python 3.11+ and is not a security risk as it only targets a built-in module.
[INDIRECT_PROMPT_INJECTION]: The skill processes local Rust source code and Cargo manifest files. While this represents a data ingestion surface, the analysis scripts only perform pattern matching and structured metadata extraction. There is no evidence of unsafe interpolation that would allow untrusted data to influence the agent's core instructions.

rust-edition2024-performance