fix: skip remote history for zero session limit

[ump45nose]

Summary

This pull request fixes OpenAIConversationsSession.get_items() when the effective session history limit is zero.

The Agents SDK session contract uses limit=0 to disable history retrieval, but this backend forwarded zero to conversations.items.list. The Conversations API only accepts limits from 1 to 100, so the call failed instead of returning no history. The method now resolves the effective limit before initializing a remote conversation and returns an empty list locally when it is zero.

This preserves the existing behavior for positive limits and None, and avoids creating an otherwise unused remote conversation.

API reference: https://developers.openai.com/api/reference/python/resources/conversations/subresources/items/methods/list

Test plan

• uv run pytest -q tests/memory/test_openai_conversations_session.py (37 passed)
• uv run ruff check src/agents/memory/openai_conversations_session.py tests/memory/test_openai_conversations_session.py
• uv run mypy src/agents/memory/openai_conversations_session.py
• git diff --check

The repository-wide verification script was not run locally; this PR remains Draft while CI supplies the broader regression signal.

Issue number

N/A

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

[seratch]

Thanks for sharing the idea here, but I don't think this additional logic for limit: 0 use cases is necessary in real-world apps.