The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/run_evals.py script executes the codex command-line utility via subprocess.run to perform automated skill evaluations. This execution is performed using safe practices, including list-based arguments and the use of the --sandbox flag.
[DYNAMIC_EXECUTION]: The evaluation runner validates Python fixtures using ast.parse and constructs temporary work environments to isolate test execution. These actions are standard for development and testing workflows.
[EXTERNAL_DOWNLOADS]: The skill configuration and evaluation files reference official repositories and issue trackers on GitHub belonging to the author (basicmachines-co). These references provide documentation and context for evaluation cases and do not involve untrusted remote code execution.
[SAFE]: The core instructions in SKILL.md promote best practices for Python development, such as strong typing, domain modeling, and local reasoning, which contribute to overall code maintainability and safety.

pythonic-code