fix: #3315 align generic dict output schemas

[Aphroq]

Summary

• Detect generic dict output types by checking their origin before rejecting non-type objects.
• Keep dict[str, int] unwrapped like bare dict when generating and validating non-strict structured output schemas.
• Add regression coverage for direct generic dict validation and wrapper-shape rejection.

Test plan

• uv run pytest tests/test_output_tool.py -k generic_dict
• bash .agents/skills/code-change-verification/scripts/run.sh

Issue number

Closes #3315

Checks

• I've added new tests (if relevant)
• I've added/updated the relevant documentation
• I've run make lint and make format
• I've made sure tests pass