architecture-fitness-function-author

Architecture fitness functions translate review findings into executable validation rules by classifying invariants as machine-checkable, human-judgment, or deferred. This planner separates automated checks from manual reviews to produce structured, testable rules from scattered repository evidence.

You can generate machine-checkable architecture rules from CodeBuddy review packets by applying the planner to classify findings into bounded, testable invariants. The Skill produces an architecture_fitness_functions.json artifact and a Markdown report with validation commands and ownership details.

Yes, architecture fitness functions drive CI gates and contract tests by producing a concrete JSON artifact and Markdown report. These outputs map validation commands and ownership details directly into downstream contract tests, CI gates, ADRs, and governance processes.

Machine-checkable invariants are automated rules evaluated locally by tools, while human-judgment invariants require manual review. The planner separates architecture findings into these categories alongside deferred automation to guide implementation priorities and structure downstream validation.

No, the fitness function planner does not mutate your repository when analyzing architecture findings. It reads repository evidence and review packets to output standalone JSON and Markdown artifacts, ensuring the source code remains unchanged during the validation planning process.