Fix: Handled AgentCardBuilder field inconsistencies

[SyedMohamedHyder]

Summary of Changes

This PR makes a few small improvements to the experimental Agent Card Builder feature:

• Updated references to AgentCard and AgentSkill fields to use snake_case instead of camelCase, in line with the project's conventions.
• Updated the doc_url field, which is not defined in the AgentCard model to documentation_url.
• Adjusted related unit tests to reflect these updates.

Rationale

Although this feature is still experimental, these changes help improve consistency and clarity:

1. Follows Model Conventions
   The models currently allow camelCase access only for backwards compatibility and issue a deprecation warning when used. These aliases are planned to be removed in version 0.3.0. Updating to snake_case ensures the code aligns with the intended usage and avoids deprecation warnings. Below is the corresponding code from a2a-sdk.

    def __getattr__(self, name: str) -> Any:
        """Allow getting attributes via their camelCase alias."""
        # Get the map and find the corresponding snake_case field name.
        field_name = type(self)._get_alias_map().get(name)  # noqa: SLF001

        if field_name:
            # An alias was used, issue a warning.
            warnings.warn(
                (
                    f"Accessing field '{name}' via its camelCase alias is deprecated and will be removed in version 0.3.0 "
                    f"Use the snake_case name '{field_name}' instead."
                ),
                DeprecationWarning,
                stacklevel=2,
            )

            # If an alias was used, retrieve the actual snake_case attribute.
            return getattr(self, field_name)

        # If it's not a known alias, it's a genuine missing attribute.
        raise AttributeError(
            f"'{type(self).__name__}' object has no attribute '{name}'"
        )

2. Avoids Potential Issues
   Since doc_url isn’t a valid field on AgentCard, replacing it with documentation_url helps prevent confusion and ensures only supported fields are used.

I have also updated the unit tests to ensure they correctly reflect the current model structure and remain reliable over time.

These are relatively minor changes, but I hope they help keep the codebase clean and aligned with the project's direction. Feedback is always welcome.

[ankursharmas]

@seanzhou1023 Please help review.

[Kunmeer-SyedMohamedHyder]

Hello @seanzhou1023! Sorry for bothering. Please have a look on this fix when you find some time!

[wukath]

can you update this comment as well?

[rohityan]

Hi @SyedMohamedHyder ,Thank you for your contribution!
Before we can proceed can you fix the failing unit tests.
You can fix the lint errors using autoformat.sh.

[rohityan]

Hi @SyedMohamedHyder , I am closing this pull request for now due to the lack of recent updates regarding the requested changes. Please feel free to reopen this (or submit a new PR) once you have had a chance to address the feedback.
Thank you for your contribution.