Multi-Agent Engineering — Operating Protocol (general MAE)

General software-engineering multi-agent pipeline. Turns a task (feature, bug fix, refactor, security or architecture change) into a verified, production-ready change by coordinating specialized agents — Planner, Builder, Optimizer, QA, Red Team, AppSec — in an iterative loop with persistent state, policy-driven security gates, and a final production decision. No agent approves its own work; a build is not "done" until the gates run and the tests pass. Use when the user asks for "the agent team", a "multi-agent pipeline", to "orchestrate this task", or wants a full analyze → plan → build → optimize → QA → security → decision cycle. Not for one-line typo fixes or purely co…