interview-mecommit 1d32b88SKILL.md · Requirements
Interview Me
Interviews the user one question at a time, each with a guess attached, until it can state their real intent with about 95% confidence.
npx skills add addyosmani/agent-skills --skill interview-meCommand selected. Copy it manually.
- Permissions
- read-repo
- write-files
- shell
- network
- Compatible with
- Claude Code, Codex, Cursor, Copilot, Gemini CLI
- GitHub
- 112.3k
- Source
- Community · Addy Osmani
- Status
- Scan passed
- License
- MIT
- Category
- Requirements
Core rule
Only an explicit yes confirms the intent; whatever you think or sounds good does not.
Execution Phases
Phase 01Hypothesize with confidence
Write your best one-sentence read of what the user wants with an honest confidence number; below about 70%, note what is missing.
Phase 02One question, one guess
Ask one focused question with your guess and its reasoning attached, and wait for the reaction before the next.
Phase 03Want vs. should want
When answers sound like best-practice talk, ask what they would actually want if they did not have to justify it to anyone.
Phase 04Restate the intent
Write back Outcome, User, Why now, Success, Constraint and Out of scope in the user's words for line-by-line confirmation.
Phase 05Confirm and stop
After an explicit yes, deliver the statement of intent, offer next steps such as saving it to docs/intent/, and end the turn.
Permissions (summarised from SKILL.md)
| Capability | Needed | Details |
|---|---|---|
| File writes | Limited | Only saves the statement of intent to docs/intent/[topic].md, and only after the user confirms. |
| Command execution | Not required | Runs no commands during the interview and starts no downstream work right after confirmation. |
| Network outbound | Not required | Conversation only; no network access. |
SKILL.md frontmatter (YAML)
---
name: interview-me
description: Extracts what the user actually wants instead of what they think they should want. Achieves this through one-question-at-a-time interview until ~95% confidence about the underlying intent. Use when an ask is underspecified ("build me X" without "for whom" or "why now"), when the user explicitly invokes ("interview me", "grill me", "are we sure?", "stress-test my thinking"), or when you catch yourself silently filling in ambiguous requirements before any plan, spec, or code exists.
---