feat: add remote fallback starter execution

[hyeokjun32]

Summary

• Added bounded fallback starter execution for remote dispatch when the primary HTTP/SSH starter fails with a retry-policy error.
• Preserved primary remote_execution_result and added fallback_execution_result only when a fallback attempt is actually made.
• Updated remote dispatch docs to clarify the fallback boundary as starter evidence, not production retry orchestration.

Validation

• PYTHONDONTWRITEBYTECODE=1 .venv/bin/python -m pytest tests/test_remote_dispatch.py -q -p no:cacheprovider
• PYTHONDONTWRITEBYTECODE=1 .venv/bin/python -m pytest -q -p no:cacheprovider
• git diff --check

Notes

• Fallback attempts are bounded by retry_policy.max_attempts and fallback_on.
• No heartbeat, auth, production failover state, or long-lived worker lifecycle management was added.