What is a grill me skill?

A skill that makes the agent question your plan, one hard question at a time, before it writes code. It looks for decisions nobody has made yet: failure handling, ownership, scale, deletion.

How does it change a session?

Without it, an agent usually starts building from your first message. With it, the agent restates the plan, asks one question at a time, tracks decisions, and writes no code until the questions are covered or you say stop.

It is set by a risk setting in your profile. The free skill asks 15 to 25 questions on Careful, 8 to 12 on Balanced and 3 to 6 on Move fast, and you can stop it at any point.

Where do the decisions go?

The skill keeps a running list during the session. An ADR skill turns each important decision into a short written record you can commit with the code.

Guide

Best Claude Code skills for planning

Most expensive bugs are decisions nobody made. A planning skill makes them early, while changing them is cheap. This is the free one with its real workflow, and the paid skills that turn answers into records and cut scope.

Start free: Grill Me

Free pick

Grill Me

FreeScan 100/100

A relentless review of your plan. Loads when the user shares a plan or asks to build something non-trivial. About 1,075 tokens when it triggers, 91 always loaded.

Its trigger text, as shipped

Use when the user shares a plan, spec, feature or architecture, asks to build something non-trivial, or says \"grill me\", \"poke holes\", \"what am I missing\". Interrogates the plan one question at a time, starting with data, failure and scale, and writes no code until the questions are resolved. Reads profile.md for project, stack and risk appetite.

How many questions it asks

## Depth

| Risk appetite | Questions | Cover |
|---|---|---|
| Careful | 15 to 25 | Every branch below |
| Balanced | 8 to 12 | Data, failure, scale, then the first three branches that apply |
| Move fast | 3 to 6 | Only decisions that are expensive to reverse (data model, public API, auth, money, deletion) |

The workflow, as shipped

## Workflow

1. Restate the plan in two sentences and ask the user to confirm. A wrong restatement is the cheapest bug you will find.
2. Ask one question per message. Never a list, never two questions joined by "and".
3. Each question is one sentence, then one line on why it matters. Name the specific scenario ("a customer downgrades mid-month"), not the category.
4. Work through the branches in this order: data, failure, scale, then flows, edge cases, security and privacy, operations.
5. After each answer, if it opens a new hole, follow it down before moving on. If it contradicts an earlier answer, say so and resolve it.
6. Keep a running list of decisions. Do not write code during the grill. If the user asks for code, offer to stop and summarise first.
7. Stop when the depth for the risk appetite is covered or the user says "enough" or "ship it".

A worked example, from the skill

## Worked example

Profile: Project Billing service, Stack Next.js and Postgres, Risk Balanced.

"Before I ask anything: I read `prisma/schema.prisma` and `app/api/webhooks`. Plan as I understand it: customers pick a plan, Stripe charges them monthly, you store the status. Right?"

User: "Yes."

"Q1 (Failure): Stripe sends the same webhook twice, or sends invoice.paid before your checkout record exists. What should happen? Why it matters: a duplicate event can double-grant access, and an early event can be dropped silently."

See Grill Me, its first lines and how to install it

What changes in a session

	Without a planning skill	With Grill Me
First move	Starts building from your first message	Restates the plan in two sentences and asks you to confirm
Questions	A few at once, or none	One per message, with one line on why it matters
Coverage	Whatever comes to mind	Data, failure, scale, then flows, edge cases, security, operations
Code	Immediately	None during the questions
End	When the agent is done	When the depth for your risk setting is covered, or you say enough

The right column is what the skill instructs, from its own text above. The left column is the common default, not a measurement of any one agent.

Paid: write it down and cut it down

Use Grill Me Pro when you want every question answered with a recommendation and a ticket list at the end. Use ADR Writer to record each decision once it is made. Use Scope Cutter when the plan is bigger than the time you have.

A plan also needs a good instructions file. See the free AGENTS.md Architect, and read skills vs CLAUDE.md vs MCP to decide where each instruction should live. For writing, see skills for writing.

Questions

What is a grill me skill?: A skill that makes the agent question your plan, one hard question at a time, before it writes code. It looks for decisions nobody has made yet: failure handling, ownership, scale, deletion.
How does it change a session?: Without it, an agent usually starts building from your first message. With it, the agent restates the plan, asks one question at a time, tracks decisions, and writes no code until the questions are covered or you say stop.
Is it slow?: It is set by a risk setting in your profile. The free skill asks 15 to 25 questions on Careful, 8 to 12 on Balanced and 3 to 6 on Move fast, and you can stop it at any point.
Where do the decisions go?: The skill keeps a running list during the session. An ADR skill turns each important decision into a short written record you can commit with the code.

Keep reading

Scan your own skill

Paste a SKILL.md or drop a folder. It is scored in your browser and nothing is uploaded.

Open the Skill Scanner

Want a skill that already passes? Start with a free one.

Grill MeFreeScan 100/100A relentless review of your plan. From the Grill Me collection.

Your bag