feat(router): add openai-compatible llm judge router

[panpan0000]

Description

This PR adds a small custom router that uses an OpenAI-compatible judge model to choose between two backend models. It also surfaces judge latency and optional routing errors in serve logs.

Related Issues

• Fixes #
• Related to #

Type of Change

• Bug fix (non-breaking change that fixes an issue)
• New feature (non-breaking change that adds functionality)
• New router implementation
• Breaking change (fix or feature that would cause existing functionality to change)
• Documentation update
• Performance improvement
• Code refactoring
• Test addition or update

Changes Made

• add llmjudgerouter as a custom router for OpenAI-compatible judge endpoints
• keep the judge output short and structured with model, confidence, reason, and signals
• add fail_on_routing_error and judge_ms logging in serve

Router Affected (if applicable)

• Other / Core framework

Checklist

• I have tested my changes locally
• I have reviewed my own code
• I have added/updated documentation where necessary
• I have added tests for my changes (if applicable)
• My changes do not introduce new warnings or errors
• Any dependent changes have been merged and published

Test Plan

python3 -m pytest tests/test_llmjudgerouter.py tests/test_llmmultiroundrouter_decomposition.py
python3 -m pytest tests/test_plugin_system.py -q

Screenshots (if applicable)

Additional Notes

• The judge router uses only the OpenAI-compatible /chat/completions path.
• The prompt is intentionally short so users can tune it for their own provider.