fix: parse Google docstrings whose Args: follows a section body

[oldrobotdev]

Summary

griffe's Google parser skips a section header when there is no blank line above it and the next line is indented, which silently drops every parameter description. #3832 normalized that for a header sitting directly below the summary, but the helper only inserts the blank line when the preceding line is non-indented. An Args: that follows another section's indented body still loses everything, and the tool reaches the model with no parameter descriptions:

def weather(city: str, units: str):
    """Get the weather for a city.

    Note:
        Results are cached.
    Args:
        city: The city to get weather for.
        units: Temperature units to use.
    """

griffe applies that gate no matter how the preceding line is indented, so this drops the non-indented restriction and both cases normalize the same way. The header regex stays anchored at column 0, so an Args: nested inside an example block still cannot match, and well-formed docstrings are still returned byte-identical.

Test plan

Two tests: parameter descriptions survive when Args: follows an indented Note: body, and that variant produces the same schema as the well-formed version with the blank line. Both fail on main and pass with the change.

.agents/skills/code-change-verification/scripts/run.sh passes (format, lint, typecheck, tests).

Checks

• I've added new tests, if relevant
• I've run .agents/skills/code-change-verification/scripts/run.sh
• I've confirmed all verification steps pass
• If using Codex, I've run /review before submitting this PR