When does the protocol decide to ask a question versus making a decision?

The protocol directs the AI to decide directly if the cost of reverting is low (like file naming) and only ask questions for high-stakes decisions like DB schemas, public API structures, or ambiguous business logic.

Why shouldn't subagents use AskUserQuestion directly?

In many CLI environments, subagent calls to AskUserQuestion are not displayed to the user. This protocol routes requests through the main context to ensure the user actually sees and can respond to the prompt.

What is the 'Think Partner' philosophy?

It is an interaction model where the AI follows the user's creative energy and provides interpretations and recommendations rather than acting like an interviewer with a rigid checklist.

Does this skill support Markdown in user prompts?

No, because CLI environments often render Markdown poorly in interactive prompts. This protocol enforces plain text formatting with clear labels and bracketed highlights for maximum readability.

Subagent Clarification Protocol

Name: Subagent Clarification Protocol
Author: Wondermove-Inc

byWondermove-Inc

•

开发者工具

Standardizes subagent-to-user communication by replacing direct CLI prompts with structured data flags for main context handling.

The Clarification Protocol is a specialized skill for Claude subagents designed to ensure seamless user interaction within CLI-based development environments. Since subagents often cannot trigger direct UI prompts like AskUserQuestion, this protocol provides a structured framework for returning decision flags and context to the main agent. It promotes a 'think partner' philosophy, encouraging the AI to follow the user's creative flow, offer concrete recommendations instead of vague options, and only interrupt the workflow for high-stakes architectural or business logic decisions.

主要功能

01Automated decision logic for determining when to act versus when to ask

021 GitHub stars

03Structured JSON-based clarification flags for subagent-to-main communication

04Think Partner philosophy to minimize user fatigue and avoid repetitive checklists

05Predefined clarification types for scope, tech decisions, and architecture patterns

06CLI-optimized plain text formatting to ensure readability in terminal environments

使用场景

01Maintaining consistent interaction patterns across different AI coding modules

02Building complex multi-agent workflows where subagents need user input

03Reducing AI-to-user questioning fatigue by prioritizing high-impact decisions

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add wondermove-inc/calab-claude-plugin clarification-protocol

For use in Claude.ai and ChatGPT

Download Skill