---

name: interactive-feedback description: Implements a proactive confirmation and continuous loop mechanism for all agent interactions. Use this skill when engaging in any task (consultation, development, debugging) to ensure step-by-step user approval and maintain continuous workflow until explicit termination.

Continuous Feedback Loop Skill

Core Mechanism

Three-Phase Loop:

1. Output Phase (IN CONVERSATION)

   • Present complete solution/plan/analysis in the main conversation thread
   • Include all relevant details: code, reasoning, alternatives, analysis

2. Confirmation Phase (IN FEEDBACK WINDOW)

   • Call mcp_mcp-feedback_interactive_feedback tool
   • Provide ONLY a one-sentence summary requesting approval
   • Example: "Approve implementing feature X with approach Y?"

3. Execution Phase (IN CONVERSATION)

   • Upon approval, execute the proposed step
   • Perform actions (write code, run commands, make changes)
   • Report execution results in the conversation

Loop Continuation:

• After completing ANY step, immediately call mcp_mcp-feedback_interactive_feedback again
• Prepare and present the next step
• Continue the loop

Termination Conditions: The loop ONLY terminates when user explicitly states:

• "任务结束" (task completed)
• "停止对话" (stop conversation)
• "结束循环" (end loop)
• Or equivalent clear termination signals

Never assume task completion without explicit confirmation.

Channel Responsibilities

Anti-Patterns

❌ Putting detailed code/analysis in feedback window
❌ Breaking loop after single task without confirmation
❌ Assuming completion without explicit user termination signal