intent-interview
One multiple-choice question at a time, until the agent knows what you actually want.
Safety scan
A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.
Check it deeper on your machine
It has changed since its scan on 21 Nov 2026.
As a Claude Code plugin, install it and then ask Claude Code:
/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/robfold/intent-interview with scanskillsafety
Or from a terminal, clone it and run it:
git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/robfold/intent-interview'
It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.
Badge for your README
Markdown
[](/listings/robfold--intent-interview)HTML
<a href="/listings/robfold--intent-interview"><img src="/api/listings/robfold--intent-interview/badge.svg" alt="findsafeskills safety rating" /></a>Install
Install instructions from the README
As written by the repo's author. findsafeskills has not verified these steps.
Claude Code:
/plugin marketplace add robfold/intent-interview /plugin install intent-interview@intent-interview
Other agents, with the [skills CLI](https://skills.sh):
npx skills add robfold/intent-interview
Or copy [`skills/intent-interview/`](skills/intent-interview/) into your agent's skills folder (`~/.claude/skills/` for Claude Code).
What was scanned
The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.
.claude-plugin/marketplace.json · Manifest · 16 lines
- {
- "$schema": "https://anthropic.com/claude-code/marketplace.schema.json",
- "name": "intent-interview",
- "description": "One multiple-choice question at a time, until the agent knows what you actually want.",
- "owner": {
- "name": "Rob"
- },
- "plugins": [
- {
- "name": "intent-interview",
- "description": "Intent interview with scene-based multiple choice and no recommended answers on intent questions.",
- "source": "./",
- "category": "productivity"
- }
- ]
- }
README.md · README · 61 lines
- # intent-interview
- English · [한국어](README.ko.md)
- An agent skill that finds out what you actually want before it builds anything — one multiple-choice question at a time.
- 
- <sub>Recorded in Claude Code with this plugin loaded. Idle time is cut and thinking is sped up. The answers follow an example scenario.</sub>
- - **One question per round.** It asks the question whose answer changes the most.
- - **Options are scenes, not feature names.** You pick a picture instead of writing a spec.
- - **No recommended answer on questions about what you want.** A recommendation pulls the answer and hides your intent.
- - **It stops on its own.** Once it can predict your next three answers, it reads the result back in your words for you to confirm.
- ## What it looks like
- > You asked for "a dashboard for the sales team". Which morning is the one you mean?
- >
- > 1. A manager opens it at 9 a.m. and sees who is behind on follow-ups.
- > 2. Each rep opens it and sees their own next five calls.
- > 3. Leadership opens it on Friday and sees the quarter's trend.
- No option is marked as recommended. Picking 2 instead of 1 changes who the user is, what data the dashboard needs, and what "done" means.
- ## Install
- Claude Code:
- ```
- /plugin marketplace add robfold/intent-interview
- /plugin install intent-interview@intent-interview
- ```
- Other agents, with the [skills CLI](https://skills.sh):
- ```
- npx skills add robfold/intent-interview
- ```
- Or copy [`skills/intent-interview/`](skills/intent-interview/) into your agent's skills folder (`~/.claude/skills/` for Claude Code).
- ## Use
- Tell your agent "interview me before you build this", or call the skill by name. It uses your harness's structured question tool when there is one (`AskUserQuestion` in Claude Code). Otherwise it shows numbered options in chat.
- The interview ends with a five-line read-back — outcome, reason, success, limits, out of scope — that you confirm before any work starts.
- ## When to use something else
- If you already have a plan and want it stress-tested, use [Matt Pocock's `grill-me`](https://github.com/mattpocock/skills). It asks several open questions per round, each with a recommended answer, and keeps clickable multiple choice [out of scope on purpose](https://github.com/mattpocock/skills/blob/main/.out-of-scope/native-question-tool.md).
- `intent-interview` is for the step before that: the request is still fuzzy, and you would rather pick than write.
- ## Credits
- Built on ideas from [addyosmani's `interview-me`](https://github.com/addyosmani/agent-skills/tree/main/skills/interview-me), [mattpocock's `grilling`](https://github.com/mattpocock/skills/tree/main/skills/productivity/grilling), and [devbrother2024's `deep-interview`](https://github.com/devbrother2024/skills/tree/main/deep-interview). The text is new. The end of [SKILL.md](skills/intent-interview/SKILL.md) lists what came from where.
- ## License
- [MIT](LICENSE)
Ratings and comments
Sign in to rate this skill.
Sign inFree to use. If it’s helped, pass it on.