The Agent Skills Directory

[SAFE]: The skill contains architectural guidelines and Python code templates for managing LLM API costs. It includes logic for model selection based on task complexity, budget tracking using immutable dataclasses, and specialized retry logic for transient network errors.
[INDIRECT_PROMPT_INJECTION]: The skill describes a pipeline that processes external text and user input. As it is primarily a documentation and design pattern resource, it does not implement risky capabilities such as file system modifications or arbitrary command execution that would be vulnerable to exploitation via indirect injection.

cost-aware-llm-pipeline