Skip to content
Back to skills

Interview

ASecurity

Use by default whenever you plan, brief, or weigh options with a person. Lead an interview that settles every open decision with structured questions, recommended answer first, and keep a living decision log until nothing material is undecided.

  • 9 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added November 10, 2026
ai-agentsgoapi

Works with

  • api

Security analysis

A100/100

Pro scans all 3 files and shows the line behind each finding

Scanned November 18, 2026

npx -y skills add greenfield-inc/agent-farm --skill interview --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Interview?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Interview
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/greenfield-inc-interview/badge)](https://www.skillsdirectory.com/skills/greenfield-inc-interview)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: interview
description: Use by default whenever you plan, brief, or weigh options with a person. Lead an interview that settles every open decision with structured questions, recommended answer first, and keep a living decision log until nothing material is undecided.
---

# Interview

You surface the open questions, ask them, and record the answers, so the person only answers.

## Find what is open

**Send a swarm of parallel explorer and research subagents** across the sources, code and comparable products to list everything that changes what gets built: scope, competing designs, limits, conflicts, permissions, assumptions. Settle what the evidence answers; bring the rest to the person.

## Ask

- **Show first.** Share the page the question is about (open it when the person has asked for that).
- **Ask in rounds.** Ask 1 to 4 questions per round, most consequential first, in the format in [references/question-format.md](references/question-format.md). Use the harness's structured question tool, unless the host has its own question mechanism.
- **When the host has its own question mechanism, its injected instructions decide the card**: an interview round whose answers go in the decision log is the host's approval card, never `AskUserQuestion`; a quick clarification that only steers your next step may use Claude's own card. Under an orchestrator, every round is a `needs_you` (standing rules, Questions for the person).
- **Recommend.** Each question has 2 to 4 real options. The recommended one goes first, marked "(Recommended)", and each option gets a one-line consequence.
- **Open-ended last.** Ask for names, numbers and dates as plain text at the end of a round.
- **Never ask for a secret.** An API key, token or password never goes in a round, a card or the decision log: get it as the standing rules' **Never ask for a secret** says (the host's secret prompt when it has one; otherwise, set up locally and confirmed ready without its value), and record only whether it is ready.
- **Clarify on confusion.** When an answer is a question or conflicts with an earlier decision, explain the trade-off and ask again. The person makes every product choice.

## Record

- Keep `decisions.html` in the bundle as in [references/decision-log-template.md](references/decision-log-template.md), rendered with `page`: one dated round per batch, reversals as new lines.
- After each round, update affected documents in place and name them.
- Under an orchestrator, running workers receive new decisions as its numbered addenda.

## Finish

Continue until every material decision is settled, then say so and list what is deferred.

Files in this skill

  • SKILL.md1.9 KB
  • references/decision-log-template.md740 B
  • references/question-format.md1.6 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

View all in ai-agents →

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.