A code smell is a recognizable symptom in source code that suggests a design, maintainability, or implementation problem. In review, smells commonly appear as poor structure, weak readability, or anti-patterns. section 3.3.3 section 13.1.2
Treat code smells as review signals, not automatic reasons to stop delivery:
- Block a pull request when the smell represents a critical maintainability problem; otherwise address it through constructive feedback, a minor update, or a separate pull request. section 3.3.3
- Use AI-powered review as an early pass to detect and refactor more smells, while reserving human attention for solution design, architecture, and project alignment. section 13.1.2
- Apply recurring feedback during development, similarly to a linter or formatter, so known issues create less review noise. section 13.1.2
Figure 3.3 Blocking vs. nonblocking issues
Use the blocking-versus-nonblocking distinction to calibrate when a code smell should stop approval. section 3.3.3
