interview

Item-by-Item Interview

Walk through a list of items one by one using AskUserQuestion, ensuring every item gets a deliberate decision.

Workflow

1. Collect Items

Identify all items from the conversation context. Consolidate if scattered. Number sequentially from 1.

2. Present Summary

**Interview: [N] items to address**
Starting walk-through. Each item will be presented for your decision.

3. Walk Through Each Item

Present each item one at a time via AskUserQuestion.

Item-to-Question Mapping

• header: index + optional severity tag — "#3 HIGH" or "#3"
• question: "[Title]: [Description] — How to address?"
• options: actionable choices, recommended first

Previews

When an item involves code or config, use preview on options to show what each action produces (proposed fix, alternative implementation, before/after). Skip preview for trivial items.

Example

AskUserQuestion({
  questions: [{
    question: "Missing null check in parse_config(): Line 42 dereferences config.timeout without None check — How to address?",
    header: "#3 HIGH",
    options: [
      {
        label: "Fix now (Recommended)",
        description: "Add null check with sensible default",
        preview: "def parse_config(path: str) -> Config:\n    config = load_yaml(path)\n    if config is None:\n        return Config(timeout=30)\n    return Config(timeout=config.timeout)"
      },
      {
        label: "Raise error",
        description: "Fail explicitly instead of defaulting",
        preview: "def parse_config(path: str) -> Config:\n    config = load_yaml(path)\n    if config is None:\n        raise ConfigError(f\"Missing: {path}\")\n    return Config(timeout=config.timeout)"
      },
      { label: "Skip", description: "Accept the risk" }
    ],
    multiSelect: false
  }]
})

4. Track Progress

After each response, acknowledge briefly and state progress:

#3 HIGH — Missing null check: **Fix now** (add null check with default)
Progress: 3/10 items addressed

5. Execute Decisions

Apply agreed actions after each item, or batch if user prefers.

6. Summarize

**Interview Complete: [N]/[N] items addressed**

| # | Severity | Item | Decision |
|---|----------|------|----------|
| 1 | HIGH | Missing null check | Fix now |
| 2 | MED | Unused import | Fix now |
| 3 | LOW | Naming convention | Skip |

Rules

• Present every item; never skip one silently.
• If the user asks to batch-fix similar items, group them but log each individually.
• If new items are discovered during fixes, append to the queue and inform the user.