Skip to content

First Ask Task Refiner

Makes the agent interview you about scope and deliverables before starting a task

Stars
40,703
Licence
MIT
Files
1 file

What it does

When you give the agent a task, it first asks you specific clarifying questions about scope, expected deliverables, constraints and success criteria, using a Joyride input tool. It can also explore your project or search the web to understand the task better. Only after you confirm there is nothing more to add does it present a plan, create a todo list and start work.

Great for

  • Developers who often give vague requests and want fewer wrong turns
  • Teams wanting a structured requirements step before an agent edits code
  • People working on large or unfamiliar tasks where upfront alignment saves rework

Skip it if

  • Quick, well-specified tasks where extra questions just slow things down
  • Environments without the Joyride extension, since the core workflow depends on its input tool and the skill will not function as designed
  • Users who want the agent to act autonomously without being interrupted, because it repeatedly pauses for input

Try asking

  • “Refactor our authentication module, but first ask me what I want changed”
  • “Plan a migration from REST to GraphQL and check with me on constraints before starting”
  • “Build a reporting feature for our dashboard; interview me about the deliverables first”
Step-by-step guide: how to use First Ask Task RefinerInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    first-ask.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip first-ask.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip first-ask.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • State your goal and any known constraints up front so the agent asks better follow-up questions
  • Answer the final 'any further input?' prompt explicitly, or the agent will keep refining
  • Check the agent's plan before it starts work, since it will proceed once you agree

What's inside

  • SKILL.md

Redistributed unmodified from github/awesome-copilot under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.