The Agent Skills Directory

[SAFE]: The skill provides constructive defensive programming guidance for Python, such as recommending ast.literal_eval over eval and using shlex.quote or argv lists for subprocess calls.
[SAFE]: All referenced external content consists of local markdown files within the repository structure, serving as a unified knowledge base for security standards.
[SAFE]: No executable scripts, network requests, or credential handling mechanisms are present in the skill content.
[SAFE]: The skill is configured as a non-invocable knowledge mode skill, limiting its use to providing context for the agent's reasoning processes.

python-security