Scoring reference

Formula

penalty(count, perIssue, cap) =
  count == 0 ? 0 : min(cap, perIssue × (1 − 0.75ⁿ) / 0.25)

score = max(0, round(
  100
  − penalty(critical, 18, 42)
  − penalty(high,     10, 28)
  − penalty(medium,    4, 14)
  − penalty(low,       1,  6)
))

The same formula runs in both the Go agent and the React frontend.

Parameters

Severity	perIssue	cap
Critical	18	42
High	10	28
Medium	4	14
Low	1	6

Diminishing returns

Each additional finding contributes less than the previous:

# findings	Marginal penalty (critical)
1st	−18 pts
2nd	−13.5 pts
3rd	−10.1 pts
4th	−7.6 pts
5th+	diminishing toward cap of −42

Score lookup table

Critical	High	Medium	Low	Score
0	0	0	0	100
1	0	0	0	72
0	1	0	0	90
1	2	3	5	64
4	30	6	31	13
10+	10+	10+	10+	0