Fix incorrect and ambiguous guardrail documentation: Input/Output Guardrails and agent sequencing

[Wania-Kazmi]

Problem:

1. The documentation states that input guardrails only run on the "first agent" and output guardrails only run on the "last agent", which might be unclear in the context of agent sequences or chains.
2. The original text for the first step of Output Guardrails incorrectly stated that it receives "the same input passed to the agent"

Solution:

1. Definitions: For "first agent" (the entry point receiving user input) and "last agent" (the one producing the final output) are added directly into the relevant notes. The text also explicitly reiterates that input guardrails apply to the first agent's input and output guardrails apply to the last agent's output.
2. Output guardrail Error: This has been corrected to reflect the actual behavior: the guardrail receives "the final output produced by the agent". This aligns the documentation with the code examples and the purpose of output guardrails.

Why this change?
Correction of output guardrail description and the clarification helps users understand exactly which agent in a potential chain will have its input or output checked by these types of guardrails, providing crucial context for system design.

[github-actions]

This PR is stale because it has been open for 10 days with no activity.

[github-actions]

This PR was closed because it has been inactive for 7 days since being marked as stale.