The Agent Skills Directory

[Subagent Data Ingestion]: The research tool accepts a topic parameter that is directly passed to a subagent. While this is the intended workflow for delegation, it represents a surface where untrusted data could influence agent behavior. Implementing clear boundary markers and input validation is a recommended practice to ensure robust execution.
[SDK Implementation]: The examples utilize a specialized ai library to handle complex streaming logic and tool result aggregation. This library-centric approach is consistent with the skill's goal of teaching advanced agentic communication patterns and shows no unusual or hidden behaviors.

ai-python-streaming-tools