grill-me

Name: grill-me
Availability: InStock
Author: forrestgrump

Interview users one question at a time and persist decisions to .agent-runs/<run-id>/decisions.md.

Updated Aug 27, 2026

Search Tags:#requirements #interview #project-kickoff #grill-me #decision-persistence #run-folder

Ready to run

npx skills add https://github.com/ForrestGrump/agent-factory-scaffold --skill grill-me-forrestgrump

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: grill-me
Source: https://github.com/ForrestGrump/agent-factory-scaffold/tree/main/.agents/skills/grill-me
Command: npx skills add https://github.com/ForrestGrump/agent-factory-scaffold --skill grill-me-forrestgrump

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Grill-me guides teams to clarify project goals by interviewing the user one question at a time and persisting decisions to the current run folder, reducing ambiguity before implementation.

Core Features & Use Cases

Structured, one-question-at-a-time interviewing flow that surfaces critical decisions in a controlled sequence.
Default recommendations and concise rationales to accelerate decision making.
File-inspection capability to inform questions and minimize unnecessary prompts by leveraging existing project context.
Durable decision logging and run-folder persistence to .agent-runs/<run-id>/decisions.md for auditability and handoff.

Quick Start

Start a grilling session by selecting a run folder (for example, .agent-runs/000-project-kickoff) and begin asking one question at a time while persisting decisions to the run folder.

Frequently Asked Questions about grill-me

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I define project scope before implementation starts?▼

To define project scope before implementation, you can use a structured interview process that asks one question at a time to surface critical decisions, provides default recommendations, and persists those decisions to a run folder for durable logging.

What's the best way to run a project kickoff interview for ambiguous design tasks?▼

Running a project kickoff interview for ambiguous design tasks works best by prompting one question at a time, leveraging file-inspection to inform questions from existing project context, and recording decisions to minimize unnecessary prompts and reduce ambiguity.

Can I use an interview-based approach to capture requirements for greenfield projects?▼

Yes, you can use an interview-based approach to capture requirements for greenfield projects, as it applies to greenfield work, feature development, and ambiguous tasks by guiding teams through a controlled sequence of questions with durable decision logging.

How do I persist project decisions for auditability and team handoff?▼

To persist project decisions for auditability and handoff, the interview process records all decisions in a decisions.md file located within the specific run folder at .agent-runs/<run-id>/, ensuring durable logging and traceability.

Do I need a current run folder to start interviewing users for project scope?▼

Yes, you need a current run folder to start interviewing users for project scope, because the process requires an active run directory to persist decisions and support inspecting existing files to inform the interview questions.

Why does the interview ask one question at a time instead of presenting a full questionnaire?▼

Asking one question at a time during the interview surfaces critical decisions in a controlled sequence, preventing user overwhelm and allowing file-inspection to inform subsequent questions, which accelerates decision making with concise rationales.

kvnxiao

linear-issues

Automates Linear issue listing, creating, updating, assigning, commenting, labeling via linear-cli with ndjson output and dry-run/id-only conventions.

Community

Intermediate