Add evaluation tooling and broaden Apple bridge coverage

[zac]

Summary

• Add deterministic evaluation tooling, baseline summaries, and CI workflow support for CodeMode evals
• Expand bridge, registry, runtime, and security support across Apple services and system UI surfaces
• Update authoring and docs to reflect the new evaluation and capability model
• Add and extend tests for bridge behavior, host validation, registry mapping, and eval runner coverage

Testing

• Updated and added unit/eval tests across bridge, runtime, and registry layers
• Deterministic eval scenarios and baseline summaries were refreshed for the new coverage
• Not run (not requested)