Agent Architect

关于

Agent Architect serves as the central intelligence hub for the ACE-core workflow, acting as a high-level advisor rather than an executor. It utilizes a rigorous classification framework to determine whether information belongs as an immutable law, a capability skill, a user workflow command, or a shared convention rule. By mandating delegation via the Skill tool, it prevents context drift and ensures that your agentic environment remains modular, scalable, and follows the single source of truth principle.

主要功能

• Strict multi-tier knowledge classification framework
• Mandatory delegation protocol to prevent direct execution
• Single source of truth enforcement for skills and rules
• Comprehensive review and validation of delegated tasks
• Architectural oversight for commands, subagents, and hooks
• 7 GitHub stars

使用场景

• Deciding the optimal storage location for new domain-specific knowledge
• Restructuring and refactoring existing Claude Code components
• Maintaining architectural integrity in complex agentic workflows