[cristianmessel-cs]

No description provided.

[cristianmessel-cs]

 ## Code Review

 **internal/review/pipeline.go:139**
 Consider avoiding a direct `fmt.Println` dependency here. Since `Pipeline` already takes a `logger`, it might be cleaner (and more testable) to inject an `io.Writer` for dry-run output via `Options`/the constructor, so tests can capture and assert on the printed content instead of relying on actual
 stdout. This also decouples the pipeline from a global process stream.

 **internal/review/pipeline.go:141**
 Returning immediately after printing the dry-run output means any code that normally runs after `PublishComment` succeeds (if present later in this function) will be skipped in dry-run mode. Worth double-checking there's no post-publish side effect (metrics, status updates, etc.) that should still execute
 even when not actually posting a comment.

 **internal/review/pipeline.go:30**
 This PR introduces branching pipeline behavior (`dryRun`) but doesn't add a test covering it. Consider adding a test that runs the pipeline with `DryRun: true` and asserts `PublishComment` is never invoked on the SCM client mock, to guard against regressions.