What does this PR do?

This PR fixes a post-refactor fallback crash in the extracted agent conversation
loop.

agent/conversation_loop.py now owns the main run_conversation() body, but the
rate-limit fallback path still referenced _pool_may_recover_from_rate_limit()
as if it were local to that module. The helper was still defined in
run_agent.py, so quota/rate-limit fallback could raise NameError before the
configured fallback provider was attempted.

The fix moves the helper into a small shared module and imports it from the
conversation loop. run_agent._pool_may_recover_from_rate_limit remains as a
compatibility alias for existing tests and external patch/import sites.

Related Issue

Related to #27719 and #27370.

Type of Change

• 🐛 Bug fix (non-breaking change that fixes an issue)
• ✨ New feature (non-breaking change that adds functionality)
• 🔒 Security fix
• 📝 Documentation update
• ✅ Tests (adding or improving test coverage)
• ♻️ Refactor (no behavior change)
• 🎯 New skill (bundled or hub)

Changes Made

• Added agent/fallback_utils.py with pool_may_recover_from_rate_limit().
• Updated agent/conversation_loop.py to import and call the shared helper
  directly in the rate-limit fallback path.
• Kept run_agent._pool_may_recover_from_rate_limit as a compatibility alias.
• Updated the Gemini/CloudCode fallback regression test to verify the
  conversation loop imports the shared helper and the old run_agent alias
  still points to the same function.

How to Test

1. Run the focused fallback tests:

   pytest -q tests/agent/test_gemini_fast_fallback.py tests/run_agent/test_provider_fallback.py -k "pool or cloudcode or fallback"

2. Run ruff on the touched files:

   python -m ruff check agent/fallback_utils.py agent/conversation_loop.py run_agent.py tests/agent/test_gemini_fast_fallback.py

3. Local results:

   29 passed in 70.91s
   All checks passed!
   

Checklist

Code

• I've read the Contributing Guide
• My commit messages follow Conventional Commits (fix(scope):, feat(scope):, etc.)
• I searched for existing PRs to make sure this isn't a duplicate
• My PR contains only changes related to this fix/feature (no unrelated commits)
• I've run pytest tests/ -q and all tests pass
• I've added tests for my changes (required for bug fixes, strongly encouraged for features)
• I've tested on my platform: WSL/Linux checkout

Documentation & Housekeeping

• I've updated relevant documentation (README, docs/, docstrings) — or N/A
• I've updated cli-config.yaml.example if I added/changed config keys — or N/A
• I've updated CONTRIBUTING.md or AGENTS.md if I changed architecture or workflows — or N/A
• I've considered cross-platform impact (Windows, macOS) per the compatibility guide — or N/A
• I've updated tool descriptions/schemas if I changed tool behavior — or N/A

Screenshots / Logs

$ pytest -q tests/agent/test_gemini_fast_fallback.py tests/run_agent/test_provider_fallback.py -k "pool or cloudcode or fallback"
.............................                                            [100%]
29 passed in 70.91s (0:01:10)

$ python -m ruff check agent/fallback_utils.py agent/conversation_loop.py run_agent.py tests/agent/test_gemini_fast_fallback.py
All checks passed!