The Proactive Scanner runs independently of PR flow. It performs scheduled repository analysis to surface trends, drift, and debt before they compound into incidents.
Scan Modes
| Mode | Frequency | What it does |
|---|---|---|
| Coverage Trends | Daily | Tracks test coverage per module over time |
| Tech Debt Scan | Weekly | Large files, circular dependencies, dead code, growing complexity |
| Convention Drift | Weekly | Does new code diverge from established patterns? Cross-team comparison |
| Security Baseline | Daily | Full SAST scan, dependency vulnerabilities |
| AI-Code Audit | Weekly | Which modules have high AI-generated code density, what's the quality |
Convention Drift Detection
Particularly valuable for mixed teams across geographies. The scanner compares coding patterns between contributors and flags divergence before it hardens into conflicting conventions.
Repository knowledge layer data:
Module: user-service
Team A commits: 89% direct service calls
Team B commits: 73% repository pattern
DRIFT DETECTED:
"user-service has diverging patterns between contributors.
Drift rate: 14% → 31% over 2 weeks."
ITS correlation: ITS for user-service trending up (4 → 7)
→ agents need more iterations = code getting harder to reason about✅ Why this matters
Rising ITS (Iterations-to-Success) in a module often correlates with convention drift. When patterns diverge, both humans and AI agents take more iterations to produce correct code.
Output
The Proactive Scanner produces three types of output:
- Report: Markdown/HTML with trends, delivered on schedule
- GitHub Issues: Auto-created for critical findings that need team action
- Metrics: JSON for dashboard integration and historical tracking