playbooks

home / skills / robzolkos / zolkos-agent-skills / interview

interview skill

/skills/interview

This skill conducts an in-depth interview about a plan and generates a detailed plan file with insights.

npx playbooks add skill robzolkos/zolkos-agent-skills --skill interview

Review the files below or copy the command above to add this skill to your agents.

Files (1)
SKILL.md
1.2 KB
---
name: interview
description: Interview me about the plan
argument-hint: [file or text]
model: opus
disable-model-invocation: true
---

The user has provided input: $1

First, determine if this input is a file path or raw text:
- If it looks like a file path (contains `/` or ends with `.md`, `.txt`, etc.), use the Read tool to read the file contents as the plan
- If it's raw text (a description, idea, or plan written directly), treat it as the plan content directly

If raw text was provided (not a file):
1. Generate a filename using the format: `PLAN-YYYY-MM-DD-<short-summary>.md` where `<short-summary>` is 2-4 lowercase words from the plan separated by hyphens (use bash `date` command to get the date)
2. Create this file in the current working directory with the initial plan text

Then interview me in detail using the AskUserQuestion tool about literally anything: technical implementation, UI & UX, concerns, tradeoffs, etc. but make sure the questions are not obvious.

Be very in-depth and continue interviewing me continually until it's complete, then write the final spec to the file (either the original file if a file was provided, or the newly created timestamped file if text was provided).

Overview

This skill interviews you in depth about a plan you provide and turns that conversation into a polished specification file. It accepts either a path to an existing plan file or raw plan text, captures detailed answers through iterative questioning, and writes the completed spec to a file. The focus is on rigorous, non-obvious questions that probe technical, UX, risk, and tradeoff details.

How this skill works

On input, the skill detects whether the text is a file path or raw plan content. If a path is supplied, it reads that file as the initial plan; if raw text is supplied, it generates a timestamped filename and stores the initial plan. It then conducts a continuous, deep interview using iterative, specific questions until all aspects are clarified, and finally writes the final specification back to the chosen file.

When to use it

  • You have a rough project idea or brief and need a full, production-ready spec.
  • You want to expand a one-paragraph plan into a detailed implementation roadmap.
  • You received a draft file and need structured follow-up questions to refine it.
  • You need to capture stakeholder decisions and tradeoffs in a persistent spec.
  • You want a persistent, timestamped record of interview-driven requirements.

Best practices

  • Provide either the raw plan text or a clear file path to start; ambiguous inputs slow the flow.
  • Expect multiple rounds of targeted questions; answer with concrete constraints, priorities, and examples.
  • Be explicit about non-functional requirements (performance, scale, security) early in the interview.
  • Flag assumptions and unknowns during answers so the skill can probe them further.
  • Review the generated filename and final spec after the interview to confirm context and naming.

Example use cases

  • Turn a short product idea into a full engineering specification with API, data, and UX requirements.
  • Refine a README or draft roadmap by asking deep implementation and deployment questions.
  • Interview stakeholders to capture acceptance criteria, edge cases, and operational considerations.
  • Create a timestamped, traceable spec file for audits or handoffs to engineering teams.
  • Explore tradeoffs between architectures (monolith vs microservices), storage options, and UX flows through targeted questioning.

FAQ

How does the skill decide if my input is a file or raw text?

It treats inputs containing path characters or common extensions (like .md, .txt) as file paths; otherwise it treats input as raw plan text.

Will it overwrite my existing files?

When given a path, it operates on that file; confirm before continuing if you do not want overwrites. For raw text it creates a new timestamped filename.