ask · buiducnhat — Claude Code skill: what it does, how to install
Halts premature coding and asks targeted clarification questions one by one before any task starts.
npx skills add buiducnhat/cobrew --skill ask -a claude-codeWhat ask does
The skill prevents the assistant from planning or jumping directly into implementation when tasks are underspecified. It begins by inspecting project context, checking docs/SUMMARY.md first and prioritizing repository coding conventions.
During clarification, the agent asks strictly one question per turn, preferably offering 2–4 predefined choices. It targets scope boundaries, success criteria, technical constraints, and non-goals. The skill forbids modifying files or generating implementation plans, restricting output strictly to focused questions and occasional dry code in the conversation.
What it helps with
- clarify ambiguous task requirements before writing code
- define project scope boundaries and explicit non-goals
- surface technical constraints early without unnecessary assumptions
- resolve conflicts between repository documentation and user requests
Example prompts for the agent
- ask I want to refactor our authentication flow
- ask Help me define requirements for the new reporting endpoint
- ask How should we structure the database export?
How to install ask
1Skills CLI (Vercel)
npx skills add buiducnhat/cobrew --skill ask -a claude-code2Manually: put the files in place
~/.claude/skills/ask/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/ask/
1Skills CLI (Vercel)
npx skills add buiducnhat/cobrew --skill ask -a codex2Manually: put the files in place
.agents/skills/ask/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add buiducnhat/cobrew --skill ask -a cursor2Manually: put the files in place
.cursor/skills/ask/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install buiducnhat/cobrew ask2Manually: put the files in place
.github/skills/ask/SKILL.md1Skills CLI (Vercel)
npx skills add buiducnhat/cobrew --skill ask -a gemini-cli2Platform CLI
gemini skills install https://github.com/buiducnhat/cobrew --path plugins/cobrew/skills/askManually: put the files in place
.devin/skills/ask/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add buiducnhat/cobrew --skill ask -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about ask
How to install ask in Claude Code?
Run npx skills add buiducnhat/cobrew --skill ask -a claude-code. Alternatively, place the file directly at ~/.claude/skills/ask/SKILL.md.
What does the ask skill do?
It guides the agent to inspect project documentation and systematically gather requirements through single, sequential questions.
Does ask allow modifying repository files?
No, the skill explicitly prohibits creating or editing files; it only allows dry code directly within the conversation messages.
Source text
SKILL.md · MIT
--- name: ask description: Structured clarification and requirements gathering through focused dialogue or with dry code. Use when a task is ambiguous, underspecified, or requires user input before any action can be taken. Do not plan or implement anything—only ask questions to collect the information needed. argument-hint: "[question]" license: MIT --- # Ask ## Workflow ### Step 1: Gather Project Context Load only the project context relevant to the current task: - If `docs/SUMMARY.md` exists, read it first. - Load only task-relevant detail docs. - Prioritize `Code Standard` docs for implementation conventions. - If docs conflict with code or user intent, use the available input/question before broad changes. Skip this step if no project docs or useful repository files exist. ### Step 2: Identify Information Gaps Determine exactly what is missing before a task can proceed: - Objective and user value - Scope boundaries and non-goals - Constraints (technical, UX, performance, timeline) - Success criteria - Key decisions with multiple valid options ### Step 3: Ask Questions (One at a Time) Ask targeted questions sequentially to close each gap. Rules: - Ask **exactly one question per message** - Prefer **multiple-choice options** when practical (2–4 choices) - Use open-ended questions only when no reasonable options exist - Do not ask questions already answered by project documentation - Do not ask about implementation details prematurely - Do not bundle multiple questions into one message ## Rules - Do not write code or modify any files, only support dry coding into output, not into artifacts - Do not produce plans, designs, or implementation artifacts - Do not make assumptions; ask instead - Keep questions short and focused - Apply YAGNI: only ask what is strictly necessary to proceed
© buiducnhat · MIT · buiducnhat/cobrew
Similar skills
All skills for Agent setup: memory, context, orchestration →ask by other authors comes first.