Scoring

CODYGIENE computes scores from 0 to 100. Higher is better.

Category scores

Each category (Security, Performance, Code Health, AI Quality) gets a score based on:

1. The number of findings
2. The severity of findings
3. The confidence of findings

Higher severity and higher confidence findings reduce the score more.

Overall score

The overall score is a weighted combination of category scores, with security findings carrying the most weight.

Score interpretation

Score	Interpretation
80–100	Healthy · few or low-severity findings
50–79	Needs attention · moderate findings
0–49	High risk · significant findings require action

These are guidelines, not thresholds. A score of 80 with one critical security finding is worse than a score of 60 with only low-severity code health findings. Always review the findings, not just the number.

PARTIAL scans

A PARTIAL scan displays a prominent warning before scores:

⚠  PARTIAL SCAN · SCORES ARE NOT REPRESENTATIVE

Scores from a PARTIAL scan are based only on analyzed files and may not represent the complete repository. Do not compare PARTIAL scan scores to COMPLETE scan scores.

Finding penalties

Each finding contributes a penalty based on:

• Severity · critical > high > medium > low
• Confidence · higher confidence = higher penalty
• Category · security findings have larger penalties

The total penalty is capped to prevent a single category from dominating the overall score.

Determinism

Scores are deterministic. The same codebase scanned three times produces identical scores, findings, fingerprints, and coverage.