Add log operation for debugging and pass-through functionality

[jmalovera10]

Description

Adds a log operation to OperationEvaluator: {"log": message} evaluates message (any
nested expression), logs it via Logger.debug/1, and returns the evaluated value unchanged
— a transparent pass-through, not a fixed sentinel. Because it returns its argument's value
rather than replacing it, log can be wrapped around any part of a rule's conditions or
actions without changing what the rule decides, purely for debugging intermediate values —
e.g. {"gt": [{"log": {"obj": "order.total"}}, 100]} logs order.total while still comparing
it normally, and a whole rule's conditions can be wrapped in log to trace its final decision.

Changes

• lib/operation_evaluator.ex: require Logger; new log clause — evaluates its argument,
  logs it via a loggable/1 helper (binaries logged as-is, everything else via inspect/1),
  and returns the evaluated value unchanged.
• test/operation_evaluator_test.exs: new describe "log operation" block (8 cases) covering
  plain-string, nested obj, and nested computed messages; non-binary stringification;
  propagated evaluation errors from a bad nested reference; composition inside a gt
  comparison; composition as a rule's entire top-level conditions (via
  StatefulRuleEngine); and wrapping set as an action.
• README.md: new "Logging Operations" table, including the pass-through note and a caution
  against logging sensitive fact values.
• CHANGELOG.md: [Unreleased] → Added entry.

Type of Change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation update
• Refactoring (no functional changes)
• Test updates

How Has This Been Tested?

• mix test — 164 tests, 0 failures, including the full pre-existing suite unmodified plus 8
  new log operation cases (plain/nested/composed messages, non-binary stringification,
  propagated errors, composition inside a comparison, composition as top-level conditions,
  wrapping set).

• mix credo — no issues.

• mix dialyzer — 0 errors.

• mix sobelow — no new findings (see Additional Notes for one pre-existing, unrelated
  line-number shift).

• Unit tests pass

• Integration tests pass

• Manual testing completed

Checklist

• My code follows the project's style guidelines
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes