audit-codebase

Define the assessment scope, baseline, and representative change scenarios, then collect evidence from contracts, source structure, version history, and runtime data. Score each simplicity principle from 0 to 3 with cited evidence and produce a structured scorecard report.

Use a diff-type assessment focused on the stated change, measuring principles like Single Responsibility, Single Knowledge Owner, and Valid States and Failures. Cite evidence as path:line references and report measured values beside each score.

No. Static metrics such as LOC, method count, complexity, coupling, and cohesion are warning signals, not verdicts. They cannot alone prove a responsibility or design problem and must be combined with structural, historical, and runtime evidence.

When no project policy exists, the default gate requires no principle scoring 0, Single Responsibility, Single Knowledge Owner, and Valid States and Failures scoring at least 2, an overall score of at least 2.2/3, and evidence coverage of at least 80%.

Do not compare raw scores across unrelated projects, since measurements are only meaningful as trends within the same system and comparable scope. Calibrate targets from the first baseline and repository trend instead of treating default thresholds as universal laws.