grilling

Interviews users in numbered rounds to stress-test plans and map decision trees.

Updated Aug 13, 2026
One-click install
npx skills add https://github.com/Evolutionary-Leadership/harness --skill grilling-evolutionary-leadership
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: grilling
Source: https://github.com/Evolutionary-Leadership/harness/tree/main/.claude/skills/grilling
Command: npx skills add https://github.com/Evolutionary-Leadership/harness --skill grilling-evolutionary-leadership

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Plans, decisions, and ideas often contain hidden assumptions and unexplored branches that only surface after work has started. This Skill stress-tests thinking by interviewing the user relentlessly until every decision in a plan is explicitly settled, producing a shared understanding before any action is taken. ## Core Features & Use Cases - Design-tree interviewing: Maps a topic as a tree of decisions and asks the whole answerable frontier in numbered rounds, with a recommended answer for every question. - Continuous question numbering: Question numbers run across rounds (Q1, Q2, ... Q7) so users can answer by number unambiguously even after scrolling back. - Grill autonomy mode: Users can delegate all remaining answers to the AI's recommendations, with mandatory breakouts for one-way decisions (schemas, migrations, public interfaces) or pure guesses. - Round reporting: Emits paired cockpit reports when a round is asked and answered, so waiting-on-human time is visible rather than looking like a stalled session. - Use Case: During /brainstorm or /feature phase 1, the skill grills the user about a proposed feature—scope, data model, edge cases—recording every decision and rejected alternative into the feature context before implementation begins. ## Quick Start Ask the AI to grill you about your plan or idea, for example: "Grill me about my plan to add team workspaces to the app."

Frequently Asked Questions about grilling

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

FAQPage Schema
How do I stress-test a plan or idea with an AI assistant?

Invoke the grilling skill by asking the AI to grill you about your plan. It maps your topic as a design tree and asks every currently answerable decision in numbered rounds, each with a recommended answer, until no assumptions remain.

How does the grilling skill structure its interview questions?

Questions are grouped into rounds containing the full frontier of decisions whose prerequisites are already settled. Each question is numbered continuously across rounds (Q1, Q2, Q7...) and includes a recommended answer, so users can reply by number.

Can I let the AI answer the grilling questions itself?

Yes, grill autonomy lets you accept the AI's recommended answers for all remaining questions of that grill. The AI still breaks out to ask you when a decision is irreversible, like a schema or migration, or when its recommendation would be a pure guess.

When should I not use a grilling-style interview?

Avoid it for tasks with no open decisions, such as mechanical execution or lookups, since the skill exists to settle branching design choices. Facts about the codebase or environment are gathered by sub-agents, not asked of the user.

How does grilling work with /brainstorm and /feature workflows?

The skill is the interview engine behind /brainstorm and phase 1 of /feature, where it reports rounds under the challenging or shaping positions. Autonomy granted in one workflow does not carry into the other, and /feature --ship grants autonomy from the first round.