Classifies user messages at process gates to automate workflow decisions for AI orchestrator agents.
The Message Routing skill provides a robust framework for AI orchestrators to interpret and act upon user input during supervised execution phases. By categorizing responses into specific buckets—Approval, Feedback, Question, or Instruction—it ensures that Claude handles user intent with high precision. This prevents the agent from proceeding prematurely when further clarification or revisions are needed, while streamlining the path to completion when clear approval is given. It is particularly useful for complex, multi-stage workflows where human-in-the-loop validation is required at critical gates.
主要功能
01Intelligent classification of user intent into four distinct categories
02Automated workflow steering based on approval signals like 'lgtm' or 'continue'
03Multi-intent processing for handling complex, compound user messages
04Structured feedback loop for logging persistent output quality issues
050 GitHub stars
06Ambiguity resolution logic to prevent incorrect automated actions
使用场景
01Automating the transition between different specialized agents based on user critique
02Standardizing feedback collection to improve AI-generated documentation and code quality
03Managing human-in-the-loop approvals during multi-stage software architecture tasks
