By scottrigby
Instructs Claude to use the AskUserQuestion tool instead of asking questions in response text
A Claude Code plugin that instructs Claude to use the AskUserQuestion tool whenever it needs to ask the user something, instead of embedding questions in plain response text.
When Claude asks questions inline in its response, the conversation flow breaks — you have to read through prose to find the question, and there's no structured way to respond. The AskUserQuestion tool creates a proper interactive prompt, keeping questions distinct from explanations and actions.
This plugin injects that instruction automatically on every prompt via a UserPromptSubmit hook, so you never have to think about it.
jq must be installed and available in your PATHAdd this repo as a marketplace, then install the plugin:
/plugin marketplace add scottrigby/claude-ask-questions
/plugin install claude-ask-questions
A UserPromptSubmit hook fires on every prompt and uses jq to inject an additionalContext instruction into the request. The instruction tells Claude to:
AskUserQuestion for any question it needs to askAskUserQuestion is a deferred tool (not yet loaded), first call ToolSearch to fetch its schema, then call itNo configuration needed — it works automatically after installation.
MIT
Based on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimnpx claudepluginhub scottrigby/claude-ask-questions --plugin claude-ask-questionsInterview-driven planning and clarification using AskUserQuestion
Example: hook-intercepted commands that run code without an API call
Collaborative technical discussion with proactive requirements gathering
Agent Workbench is an interactive UI that enhances Claude Code with rich visual components. WORKFLOW: 1) `awb run --title "Title" <component> [args]` sends an interaction and returns an ID, 2) `awb wait <id>` blocks until user responds. COMPONENTS: confirm (yes/no prompts before destructive actions), progress (multi-step task tracking with --steps), diff (code review before commits), table (structured data display), code (syntax-highlighted file viewer), json (formatted JSON/API responses), ask (quick user questions), chart (data visualization). WHEN TO USE: Before deleting files, deploying, or destructive operations use confirm. For multi-step tasks use progress. Before git commits show diff. For structured output use table/json. Always include --title. Run `awb --help` or `awb run --help` for details. The user runs `awb ui` to open the playground and respond to interactions.
Intelligent prompt optimization: injects the right context at the right moment so Claude lands a better first output. Clarifies vague prompts with research-based questions, plus targeted nudges for approach selection, plan readability, workflow routing, background execution, subagent routing, output readability, user-decision questions, and plan-mode assessment
Forces skill evaluation before every response