https://nmn.gl/blog/ai-senior-developer
and from the comments https://nmn.gl/blog/hn-rank-1-analysis

The code analyzer, analyzing linearly, often got stuck in details. To improve the analysis, we changed the approach, modeling the Mindset of experienced developers:

📝 File grouping: files are grouped by functionality (e.g., "authentication", "database").

ℹ️ Context: a description of the group's functionality within the overall architecture is added before code analysis. Impact analysis: consider changes in relation to the entire system

🕰 Historical understanding: track why the code evolved in a certain way

prompt to the group

Analyzing authentication system files:
- Core token validation logic
- Session management
- Related middleware

Focus on:
1. How these integrate with existing auth patterns
2. Security implications
3. Performance impact on other systems

Files to analyze:
${formatFiles(group.files)}

The result was an improvement in the quality of understanding, from simple observations to identifying potential problems, such as conflicts and relationships between components.

#promts