grilling

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.

Work the tree in rounds. The frontier is every decision whose prerequisites are already settled: the questions you can ask now without guessing at answers you haven't heard yet. Ask the whole frontier in one round: number each question and give your recommended answer. Then wait for the user's answers before the next round.

Format a round like so:

❓ **Q1** - **<question title>**: <question body, might be multiple paragraphs, including multiple choices>

➡️ <your recommended answer>

---

❓ **Q2** - **<question title>**: <question body, might be multiple paragraphs, including multiple choices>

➡️ <your recommended answer>

Word each question so "yes" accepts your recommended answer — except where the user has said a decision is theirs to make (a method, a baseline, a framing). There, lay the options out neutrally with the tradeoffs, recommend if asked, and don't pre-load the answer.

Each round the user answers reshapes the tree: settled decisions push the frontier outward and unblock questions that depended on them. Recompute the frontier and ask the next round. A question whose answer depends on another question still open in this round belongs to a later round, not this one. A calling skill may fix the cadence (one question per round, a set order); honor it.

Finding facts is your job, never the user's. When a frontier question needs a fact from the environment (the codebase, the filesystem, a tool), dispatch a sub-agent to find it; don't ask the user for anything you could look up yourself. Don't block on it: a running exploration is an unsettled prerequisite, so only the questions downstream of it wait for the sub-agent to report; ask the rest of the frontier now. The decisions are the user's: put each to them and wait.

The session is done when the frontier is empty: every branch of the design tree visited, nothing left silently assumed. Do not act on it until the user confirms you have reached a shared understanding.

When to grill proactively

"If in doubt" is too weak a trigger — an agent handed a concrete task rarely feels in doubt, even when it hides consequential decisions. Grill before acting, not only when asked, whenever a step would:

• change or restart a running service or process (a training run, a server, a scheduled job);
• allocate a shared resource — GPUs, ports, memory, disk, a rate-limited API quota;
• edit an interdependent config whose blast radius you cannot fully see;
• invalidate or contradict a decision already made this session.

In these cases, surface the conflict explicitly and grill on how to resolve it before proceeding.