About this skill

From fuzzy ideas to decidable structure

Grill Me Request Clarity targets a common engineering collaboration problem: a request starts as “build this feature,” “why does it break,” or “this code is messy,” while the real work depends on main flows, edge cases, exception branches, tradeoffs, and risk tolerance. Jumping straight into code or a task list can miss critical paths. The skill positions the model as a rigorous requirement interrogator, aiming not to generate a todo list but to make each decision-tree branch explicitly resolvable.

How it works and where it applies

It first extracts the core intent and classifies the request as requirement analysis, design, bug fixing, refactoring, or investigation, loading relevant templates when categories overlap. Rather than firing off questions, it makes a preliminary breakdown, marks information gaps, then asks only 3-5 high-impact questions per round, prioritizing the happy path before exceptions, boundaries, and tradeoffs. The output can use branch documents, comparison matrices, hypothesis-verification trees, or scenario tables, clearly marking confirmed items, open questions, and risks. It fits requirement clarification, solution design, incident triage, and technical-debt discussion; it does not replace formal acceptance, security audits, or legal judgment. For simple requests, it should stay concise instead of forcing a full process.

Screenshots

Use Cases

After receiving a vague “build a refund flow” request, break down main steps, edge states, and exception branches into a review-ready document.
While investigating intermittent timeouts, list hypotheses, validation paths, log signals, and recovery options before changing code.
Before refactoring, assess module ownership, dependency boundaries, and technical-debt risks to recommend keep, replace, or split actions.
When designing an API, compare sync/async, rate limiting, and idempotency tradeoffs, clearly marking confirmed and open decisions.

Best For

Product engineers reviewing features: turn a fuzzy request into main paths, exception branches, and open questions.
Backend engineers taking over legacy systems: clarify module ownership, dependency boundaries, and risks before refactoring.
Server engineers handling intermittent incidents: organize troubleshooting through hypothesis-and-validation paths.
Architects reviewing technical designs: compare async, idempotency, and rate-limiting tradeoffs while marking risks.