AI Agent Hub
Back to skills
Grill Pro Relentless Design Interview icon

Grill Pro Relentless Design Interview

AI Agent Updated 2026.08.30

Paste the following prompt into your AI chat to install this skill:

Please install @user_dc515faf/grill-stan according to https://skillhub.cn/install/skillhub.md.

About this skill

Problem

Design reviews often suffer from drifting terminology, fuzzy boundaries, and stated assumptions that disagree with the code. Engineers may use “account,” “cancellation,” or “order” differently across conversations, and those mismatches reappear later as rework. Grill Pro treats the plan as something to be tested rather than accepted, using continuous questioning to surface unclear premises.

How It Works

It walks the design tree one question at a time, resolving dependencies before moving deeper. If a question can be answered by inspecting the repository, it explores the codebase instead of asking the user. If the code contradicts a stated assumption, it calls the conflict out directly: “The code cancels the whole Order, but you said partial cancellation is possible — which behavior is intended?”

During the session it performs several checks:
- Terminology alignment: flags conflicts against the domain glossary in CONTEXT.md.
- Boundary stress-testing: invents concrete scenarios to expose edge cases around ambiguous concepts.
- Code cross-referencing: maps user claims back to actual behavior and surfaces inconsistencies.
- Inline documentation: updates CONTEXT.md as terms are resolved, keeping it to project-specific nouns rather than implementation details.
- Sparingly proposed ADRs: suggests docs/adr/ entries only when a decision is hard to reverse, surprising without context, and the result of a real trade-off.

Fit and Limits

This skill fits architecture reviews, domain modeling, API boundaries, migration plans, and other discussions that need shared understanding. It is less useful for one-off scripts, copywriting, or tasks with no code context. If no repository is available, it can still run the interview but skips CONTEXT.md and ADR updates. Teams with stricter documentation standards should align the generated file format and ADR naming with existing conventions.

Use Cases

  • Before refactoring order cancellation, align user intent, edge cases, and existing code behavior.
  • While reviewing the payment account model, clarify Customer, User, and Account boundaries and record them in CONTEXT.md.
  • Before migrating inventory service, stress-test oversell, rollback, and idempotency boundaries against the code.
  • When designing an approval workflow, identify hard-to-reverse trade-offs before writing an ADR.

Best For

  • Order domain engineers who need to align cancellation and refund boundaries before API changes.
  • Product engineers leading domain modeling who need to converge vague concepts into project-specific terms.
  • Tech leads planning service migrations who need to stress-test legacy behavior and decide on ADRs.
  • Backend engineers reviewing API contracts who need to verify interface semantics against code behavior.