A Codex Skill that manages the full lifecycle of open-source AI model adaptation under explicit authorization gates — plan, authorize, start, watch, pause/resume, stop, blind-reviewed evaluation, export, and database updates — with every safety gate enforced in code, not in prompts.
windows model-evaluation ai-safety ai-agents mlops model-adaptation campaign-orchestration codex-skill local-first-ai authorization-gates
-
Updated
Aug 18, 2026 - Python