Last released Aug 5, 2026
Deterministic guardrails for AI-written code: diff budget, scope discipline, test integrity, dependency policy — proven against an executable red-team corpus.
Supported by
