eliniscan:scan

CRITICAL: Before doing ANYTHING, you MUST ask the setup questions below. Do NOT start scanning until all questions are answered. Do NOT assume defaults — ASK the user.

<execution_context> @$HOME/.claude/eliniscan/workflows/scan.md </execution_context>

You MUST use AskUserQuestion or direct questions to get answers for ALL of these. Do NOT skip any. Do NOT assume defaults. Wait for user answers.

Display this header first:

═══════════════════════════════════════════════════
  eliniscan — Full Codebase Scanner
═══════════════════════════════════════════════════

Then ask these questions ONE BY ONE:

Question 1 — Scan Depth:

Scan depth?
  1. Full  — Read every line of every file (thorough, slower)
  2. Quick — Focus on critical patterns only (faster)

Question 2 — AI Model:

Which model should scan your files?
  1. Opus   — Most thorough, catches subtle issues (slower)
  2. Sonnet — Balanced speed and quality (recommended)
  3. Haiku  — Fastest, may miss subtle issues

Question 3 — File Types:

Which file extensions to scan? (comma-separated)
  Default: ts,tsx,js,jsx,css
  Enter custom or press enter for default:

Question 4 — Exclude Directories:

Directories to exclude? (comma-separated)
  Default: node_modules,.next,dist,build,.git
  Enter additional or press enter for default:

Question 5 — Severity Filter:

Minimum severity to report?
  1. All      — Report everything (CRITICAL → INFO)
  2. High     — Only CRITICAL and HIGH
  3. Critical — Only CRITICAL

After ALL questions are answered, display a summary:

  Scan Config:
  ─────────────────────────────
  Depth:      {depth}
  Model:      {model}
  File types: {types}
  Exclude:    {dirs}
  Severity:   {severity}

  Starting scan...