The Agent Skills Directory

[SAFE]: The skill acts as a documentation guide for the Modal framework, which is a well-known service for infrastructure and machine learning. All demonstrated code patterns, including @modal.enter, @modal.exit, and @modal.parameter, are standard features of the platform used for efficient resource management.
[PROMPT_INJECTION]: The skill demonstrates patterns where a predict method accepts a prompt string. This creates a surface for indirect prompt injection if the model being served is a large language model. While no malicious instructions are present, the examples do not demonstrate input sanitization or boundary markers, which is standard for architectural documentation but worth noting as a potential attack surface in implementation.

modal-classes-lifecycle