Add a compare tool to run analysis across two providers

[iiivanova]

Summary

Add a compare() tool that runs the same analysis through two providers and returns both outputs side by side.

Motivation

Bridgekit supports multiple providers but there's no easy way to see how they differ on the same input. A compare tool makes that a one-liner and is useful for anyone evaluating which model works best for their use case.

Expected usage

from bridgekit import compare

text = """
[analysis writeup]
"""

print(compare(text, tool="evaluate", providers=["anthropic", "openai"]))

Parameters:

• tool - which tool to run: "evaluate", "plan", "redteam"
• providers - list of two providers to compare (defaults to ["anthropic", "openai"])
• model_a, model_b - optional model overrides for each provider

Implementation notes

• Single public function in bridgekit/compare.py, exported from __init__.py
• Calls the specified tool twice with different providers and formats output side by side
• Add tests under tests/test_compare.py
• Document in README