interview-framework

# Interview Framework

Canonical interview algorithm for all spec phases. Each phase references this skill instead of duplicating the ~50-line algorithm.

## Option Limit Rule

Each question MUST have 2-4 options (max 4 for better UX). Keep most relevant options, combine similar ones.

## Single-Question Loop Structure

```text
Initialize:
  askedCount = 0
  responses = {}
  intent = [from .progress.md Intent Classification]
  minRequired = intent.minQuestions (phase-adjusted)
  maxAllowed = intent.maxQuestions (phase-adjusted)
  completionSignals = ["done", "proceed", "skip", "enough", "that's all", "continue", "next"]

Loop:
  WHILE askedCount < maxAllowed:
    |
    +-- Select next question from phase-specific pool
    |
    +-- Apply question piping: replace {var} with values from .progress.md
    |
    +-- Check parameter chain: does answer exist in .progress.md?
    |   |
    |   +-- Yes: SKIP this question, continue to next
    |   |       Log: "Skipping [question] - already answered in previous phase"
    |   +-- No: Proceed to ask
    |
    +-- Ask single question:
    |   AskUserQuestion:
    |     question: "[Current question text with piped values]"
    |     options:
    |       - "[Option 1]"
    |       - "[Option 2]"
    |       - "[Option 3]"
    |       - "Other"
    |
    +-- Store response in responses[questionKey]
    |
    +-- askedCount++
    |
    +-- Check completion conditions:
    |   |
    |   +-- If askedCount >= minRequired AND user response matches completionSignal:
    |   |   -> EXIT loop (user signaled done)
    |   |
    |   +-- If askedCount >= minRequired AND currentQuestion == finalQuestion:
    |   |   -> EXIT loop (reached final optional question)
    |   |
    |   +-- If user selected "Other":
    |   |   -> Ask follow-up (see Adaptive Depth)
    |   |   -> DO NOT increment toward maxAllowed
    |   |
    |   +-- Otherwise:
    |       -> CONTINUE to next question
```

## Completion Signal Detection

After each response, chec