Interactive Clarify

Interactive Clarify

A universal 'Ask User Questions' MCP tool for AI coding agents.

Category
Visit Server

README

Interactive Clarify

A universal "Ask User Questions" MCP tool for AI coding agents. When an AI agent encounters ambiguity, it can open structured clarifying questions in either a VS Code webview panel or a browser UI.

Works with Codex CLI, Claude Code, Factory Droid, and any MCP-compatible agent.

Architecture

AI Agent (Codex / Claude Code / Factory Droid)
  │  MCP stdio
  ▼
MCP Server (Node.js)
  │  Auto / VS Code only / Browser only
  ├──→ VS Code Extension ──→ Webview Panel
  └──→ Local Browser Window
  ▼
Returns JSON answers to agent

Features

  • Tab-based navigation — Questions shown as tabs, with click and keyboard navigation
  • Rich options — Each option has a label, description, and optional preview content
  • Recommended path — The first option is visually marked as recommended, with rationale shown in preview
  • Single & multi-select — Support for both exclusive and multi-choice questions
  • Freeform answer path — Every question supports a selectable freeform answer, and multi-select questions can combine checked options with freeform text
  • Per-answer notes — Selected structured answers can carry optional notes from the preview panel
  • Partial submit — Submission is allowed even when not every question is answered
  • Draft restoration — In-progress answers survive reloads for the same question set
  • Late-submit recovery — If the live MCP requester times out or disconnects, VS Code can still save the submitted response for later retrieval
  • VS Code themed — Webview uses native VS Code CSS variables for dark/light theme
  • Browser fallback — Opens the same React UI in your default browser when VS Code isn't available
  • Universal MCP — Any MCP-compatible agent can use it

Available MCP Tools

  • interactive_clarify
    • default tool
    • prefers VS Code, falls back to browser
    • supports mixed single-select and multi-select question lists in one request
  • interactive_clarify_vscode
    • forces the VS Code extension UI
    • returns an error if VS Code is unavailable
  • interactive_clarify_browser
    • forces the browser UI
    • shows the local browser URL in the page itself
  • interactive_clarify_get_late_response
    • retrieves a response saved after the original live MCP request timed out or disconnected

Project Structure

packages/
├── shared/              # Shared types, constants, and IPC protocol
├── mcp-server/          # MCP server (stdio transport) + browser fallback
└── vscode-extension/    # VS Code extension + React webview panel
    └── src/webview/panel/

Quick Start

Prerequisites

  • Node.js 18+
  • pnpm 9+
  • VS Code (optional, for in-editor webview UI)

1. Build

git clone <repo-url>
cd ask-user-questions-tool-interface
pnpm install
pnpm -r build

2. Configure your AI agent

Pick the agent you use and add the MCP server config (see Agent Configuration below).

3. (Optional) Install VS Code Extension

cd packages/vscode-extension
pnpm run package          # creates .vsix file
code --install-extension interactive-clarify-vscode-0.1.0.vsix

Note: pnpm run package uses vsce which is included as a devDependency. If you see vsce: command not found, make sure you've run pnpm install first.

The package script already passes --no-dependencies because vsce runs npm list internally, which conflicts with pnpm workspaces. If you need to run it manually:

npx @vscode/vsce package --no-dependencies

The extension activates automatically on VS Code startup and listens for incoming questions on a Unix socket.

4. Test it

Without VS Code (browser UI):

# With Codex CLI
codex "Use the interactive_clarify tool to ask me which database I prefer — PostgreSQL, MySQL, or SQLite — before suggesting an implementation"

The questions will open in your default browser using the same React UI.

With VS Code extension installed:

Run the same command — questions will appear in a VS Code webview panel instead of the browser.

Each question controls its own selection mode. Set multiSelect: true on the questions that should render as checkbox groups, and omit it or set false on the ones that should render as radio groups. Multi-select questions keep Freeform chat as an additional answer path alongside checked options.

Smoke test (no agent needed):

Pipe raw JSON-RPC to the MCP server to verify it works:

echo '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"capabilities":{},"clientInfo":{"name":"test","version":"0.1.0"},"protocolVersion":"2024-11-05"}}
{"jsonrpc":"2.0","method":"notifications/initialized"}
{"jsonrpc":"2.0","id":2,"method":"tools/call","params":{"name":"interactive_clarify","arguments":{"questions":[{"question":"Which database should we use?","header":"Database","options":[{"label":"PostgreSQL","description":"Relational DB with extensions"},{"label":"SQLite","description":"Lightweight file-based DB"}],"multiSelect":false}]}}}' | node packages/mcp-server/dist/bin/ask-user-mcp.js

This opens the browser fallback so you can verify the tab-based question UI works.

Agent Configuration

Codex CLI / Codex App

Add to ~/.codex/config.toml:

[mcp_servers.interactive-clarify]
command = "node"
args = ["/absolute/path/to/packages/mcp-server/dist/bin/ask-user-mcp.js"]

Tip: This config file is shared between Codex CLI and the Codex VS Code extension (Codex App). Configuring it once makes the tool available in both.

To instruct Codex to use the tool proactively, add to your system prompt or project instructions:

When you encounter ambiguous requirements or have clarifying questions,
use the interactive_clarify tool to ask the user before proceeding.

Claude Code

Add to .mcp.json (project-scoped, committed to git) or ~/.claude.json (global):

{
  "mcpServers": {
    "interactive-clarify": {
      "command": "node",
      "args": ["/absolute/path/to/packages/mcp-server/dist/bin/ask-user-mcp.js"]
    }
  }
}

Or via CLI:

claude mcp add interactive-clarify node /absolute/path/to/packages/mcp-server/dist/bin/ask-user-mcp.js

Note: Claude Code already has a built-in AskUserQuestion tool. The interactive_clarify tool name avoids any conflict. Both tools can coexist — Claude will use whichever is more appropriate.

Factory Droid

In an active Droid session:

/mcp add interactive-clarify node /absolute/path/to/packages/mcp-server/dist/bin/ask-user-mcp.js

Note: Factory Droid also has a built-in ask_user tool. The interactive_clarify name avoids conflicts.

Generic MCP Client

Any MCP-compatible client using stdio transport:

{
  "command": "node",
  "args": ["/absolute/path/to/packages/mcp-server/dist/bin/ask-user-mcp.js"]
}

If you install the server package globally or invoke package binaries directly, the published bin name is interactive-clarify-mcp.

Reference

Detailed reference is in docs/reference.md, including:

  • tool schemas
  • runtime flow
  • troubleshooting
  • keyboard behavior and answer semantics
  • package responsibilities and development entry points

Environment Variables

Variable Description
INTERACTIVE_CLARIFY_SOCKET Override the default IPC socket path. Default: ~/.interactive-clarify/interactive-clarify.sock

Development

# Watch mode for all packages
pnpm dev

# Rebuild everything
pnpm -r build

# Run the MCP server directly
pnpm --filter @interactive-clarify/mcp-server start

# Package VS Code extension
cd packages/vscode-extension
pnpm run package

License

MIT

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured