feat(evaluators): put runtime records behind a flag

[namrataghadi-galileo]

Summary

• Gate structured Galileo Luna scorer-invoke data behind GALILEO_FEATURE_FLAG_SCORER_INVOKE_RUNTIME.
• Only the exact value enabled sends the serialized Galileo record and runtime execution context; an unset or other value preserves the legacy request and its inputs.

Scope

• User-facing/API changes: None.
• Internal changes: Read the feature flag within the Galileo evaluator package and gate record generation and execution-context serialization in its Luna client. Existing scorer identity, authentication, timeout, and response handling remain intact.
• Out of scope: Generic Agent Control code, Orbit calls, and authentication-flow changes.

Risk and Rollout

• Risk level: low.
• Rollback plan: Unset the flag or set it to a value other than enabled to restore legacy request behavior.

Testing

• Added or updated automated tests for missing, disabled, non-exact, and enabled flag behavior.
• Ran Galileo package tests (make test): 205 passed.
• Ran Galileo package lint (make lint) and typecheck (make typecheck).
• Ran make check (individual package test, lint, and typecheck targets were run instead).
• Manually verified behavior.

Checklist

• Linked issue/spec (if applicable).
• Updated docs/examples for user-facing changes.
• Included any required follow-up tasks.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!