Intent Gathering

• User explicitly asks for cost estimates, service recommendations, or code analysis
• Empty workspace (prereq’s zero-code-path handles it)
• Code exists but no infra files (no .bicep, .tf, azure.yaml, or infra/ dir) — full pipeline is the only sensible path
• Intentionally vulnerable app signals (🛑 detection) — proceed to prereq directly

If azd template detected

⛔ You MUST read azd-template-routing.md for detection criteria, gate presentation, and routing protocol.

If NO azd template BUT infra files present

Ask ONE ask_user question:

1. Yes — analyze and deploy end-to-end (Recommended)
2. Just scaffold Bicep/Terraform
3. Just deploy it (I have IaC ready)
4. Other

Option 1 / vague → full pipeline (Step 3). Any other → invoke {"skill": "azure-prepare"}.

After Prereq Returns (Step 4 — scan-informed intent gathering)

Prereq has written prereq-output.json and context.json.components[] — this is the authoritative source for all downstream phases (prepare and scaffold consume context.json, not prereq-output.json).

Confirm the Azure target (“☁️ Azure target: {subscriptionName} ({subscriptionId})”). If the user wants a different subscription, write to context.json.overrides[].

Present scan results first, then ask only what prereq didn’t answer (≤2 if mostly covered, ≤4 if gaps):

⛔ Do NOT ask about stack/language (always answered by prereq) or budget (only if user mentioned cost). User corrections → context.json.overrides[]. Stop when covered or user says “just go.”

Update intent: Merge scan results into context.json.intent. Set refinedFromScan: true and populate scanDiscoveredFacts[].