Your subscription has insufficient quota for this resource when creating a new environment and trying to provision hosted agent

## Goal
Run an agent in a different subscription, region, model, etc.

## Context

The `init` flow guides you through selection of a subscription, region, and defaults new resource group, and allows you to pick a model or deploy a new one. However, `azd env new` creates a new environment and doesn't guide you through selecting all the necessary configs like `init` does which can result in errors.

## Current Behavior

I started a new project with `init` and picked my subscription, region, model, etc. That worked.

Then I need to test it in another subscription, region, etc. so I created a new env with `azd env new `. I ran `azd up` and it asks me to pick my subscription and region - but it does not have me go through the same flow of verifying the model / quota is still valid in that new selection so I end up with:

```
(x) Failed: Validating deployment

ERROR: Your subscription has insufficient quota for this resource.

Suggestion: Check current usage with 'az vm list-usage --location ' or request a quota increase in the Azure portal.

• Increase Azure subscription quotas

error executing step command 'provision': deployment failed: error deploying infrastructure: validating deployment to subscription:

Validation Error Details:
InvalidTemplateDeployment: The template deployment '' is not valid according to the validation procedure. The tracking id is '317d1c7e-fb7f-4c52-915b-6a8195232acd'. See inner errors for details.
InsufficientQuota: Insufficient quota. Cannot create/update/move resource 'ai-account-'.
```

## Desired Behavior

I would be prompted and able to provide new values for the other options like a model too. However, we currently put that in the `deployment` section of the azure.yaml, so its not a per-environment setting.

## Implementation Notes (optional but powerful)

To make this per-environment, follow the advice below: "write the deployment section of azure.yaml using expandable_string (which can be mapped to env vars) and write the values to the azd env. That way, in a new env, you would re-prompt for the values".

## Acceptance Criteria
- [ ] Unit tests added or updated
- [ ] Existing test suite passes
- [ ] Lint or format passes

## How to Test
1. `azd ai agent init -m "https://[raw.githubusercontent.com/microsoft-foundry/foundry-samples/refs/heads/main/samples/python/hosted-agents/agent-framework/echo-agent/agent.yaml](https://raw.githubusercontent.com/microsoft-foundry/foundry-samples/refs/heads/main/samples/python/hosted-agents/agent-framework/echo-agent/agent.yaml)"
2. `azd up` - pick a model, etc.
3. `azd env new`
4. `azd up` pick a different model, etc.