Multi-Pass Review Architecture

Multi-pass review splits large reviews into focused passes to avoid attention dilution and contradictory findings.

The Attention Dilution Problem

Reviewing a large codebase (14+ files) in one pass causes the model to miss issues in the middle of long contexts. Findings may also contradict each other when the model loses track of cross-file dependencies.

Multi-Pass Architecture

Pass 1 -- Per-File Local Analysis:

• Each file gets focused, independent analysis
• Catches local issues: bugs, style violations, security problems
• Each pass has the model's full attention on one file

Pass 2 -- Cross-File Integration:

• Separate pass focused on interactions between files
• Catches: data flow issues, API contract mismatches, dependency problems
• Uses summaries from Pass 1 as context

Benefits

1. No attention dilution: Each file gets full attention
2. No contradictions: Cross-file issues resolved in integration pass
3. Parallelizable: Per-file passes can run concurrently
4. Scalable: Works for any number of files

Combined with Independent Review

For maximum quality:

1. Generator instance produces code
2. Per-file review passes catch local issues
3. Cross-file integration pass catches interactions
4. All review passes use independent instances (no generator context)