akm 0.7.0 Secures AI Agent Workflows with New Proposal Review System

The release of akm 0.7.0 introduces a critical Proposal Queue that separates AI generation from direct execution, effectively solving the trust gap in agent-driven development. By integrating the akm-bench framework and opt-in security gates, it provides a robust infrastructure for developers to measure and validate AI-suggested improvements safely.