
Deep Interview
Asks one question at a time to turn a vague idea into a clear, approved spec before work starts
- Repository
- Yeachan-Heo/oh-my-claudecode
- Stars
- 40,512
- Licence
- MIT
- Files
- 1 file
What it does
When you give the agent a vague idea, this skill makes it interview you with targeted Socratic questions that expose hidden assumptions. It scores how clear the requirements are after each answer and refuses to start building until the score passes a threshold and you explicitly approve a scoped plan. The result is a written spec file (saved under .omc/specs) rather than code changes.
Great for
- Developers starting a complex feature who keep getting results that miss what they meant
- People who know they have a fuzzy idea and want the agent to ask before assuming
- Teams wanting a written, approved spec before autonomous execution begins
Skip it if
- Detailed requests that already include file paths and acceptance criteria, since the interview adds delay without benefit
- Quick single-line fixes, where the questioning overhead is wasteful
- Users who want fast autonomous output without answering questions, because the skill is designed to pause for approval
Try asking
- “deep interview me about adding user authentication to this app”
- “I have a vague idea for a dashboard, ask me everything before planning”
- “Don't assume anything, make sure you understand what I want for the export feature”
- “--frontier interview me on the new billing workflow”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 1 original files, the MIT licence and an attribution note.
deep-interview.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip deep-interview.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip deep-interview.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 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
- Answer each question in your own words; the skill is built to ask one at a time, so short answers keep the loop moving
- Check the ambiguity score it displays and the weakest dimension it names, so you know why it is asking what it asks
- Approve the scoped execution path explicitly only when the spec matches what you want
What's inside
- SKILL.md
Redistributed unmodified from Yeachan-Heo/oh-my-claudecode 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.