domains / general / plan-stress-testing / grilling

grilling

Skill card

combined

The two copies describe the same 'grilling' workflow but prescribe fundamentally different interaction patterns. The mattpocock/skills version mandates a 'rounds' and 'frontier' approach where all currently answerable questions are asked at once. The sickn33/agentic-awesome-skills version explicitly forbids this, stating 'Asking multiple questions at once is bewildering' and requiring questions to be asked 'one at a time'. Because they disagree on the core execution mechanic, crowning one would hide a valid alternative approach to the same task.

Repositories

Other copies

Combined skill

---
name: grilling
description: Grill the user relentlessly about a plan, decision, or idea. Use when the user wants to stress-test their thinking, or uses any 'grill' trigger phrases.
---

Interview the user relentlessly until you reach a shared understanding. Map this as a **design tree**: every decision branches into the decisions that hang off it.

**Interaction Mode:**
You may choose one of two modes based on user preference or context:
1. **Round-Based (Frontier):** Ask all questions whose prerequisites are settled in a single round. Number them and provide recommended answers. Wait for all answers before the next round.
2. **Sequential (One-by-One):** Ask one question at a time. Wait for the user's answer before asking the next. This is often less overwhelming for complex topics.

**Process:**
1. Identify the **frontier**: decisions whose prerequisites are already settled.
2. For each question in the current set (whether one or a round):
   - Word the question so "yes" accepts your recommended answer.
   - Provide your recommended answer clearly.
3. **Fact-Finding:** If a question requires a fact from the environment (filesystem, codebase, tools), dispatch a sub-agent or explore the codebase yourself. Do not ask the user for facts you can look up. Do not block the entire session on a single fact lookup; ask other independent questions while waiting.
4. **Update Tree:** As the user answers, update the design tree. Settled decisions unblock new questions.
5. **Completion:** The session is done when the frontier is empty (no more questions can be asked without new information) and the user confirms shared understanding.

**Format (Round-Based Example):**
```
❓ **Q1** - **<question title>**: <question body>
➡️ <your recommended answer>

---

❓ **Q2** - **<question title>**: <question body>
➡️ <your recommended answer>
```

**Format (Sequential Example):**
```
❓ **Q1** - **<question title>**: <question body>
➡️ <your recommended answer>

[Wait for user response before proceeding to Q2]
```