Execution failed: Error: Failed to get response from the AI model; retried 5 times

### Describe the bug

I have the following issue in GitHub Copilot CLI. It happened using Opus 4.8, and it unfortunately used 15 PRs without performing anything meaningful.

```
● Request failed due to a transient API error. Retrying...

● Request failed due to a transient API error. Retrying...

● Request failed due to a transient API error. Retrying...

● Request failed due to a transient API error. Retrying...

● Request failed due to a transient API error. Retrying...

Execution failed: Error: Failed to get response from the AI model; retried 5 times (total retry wait time: 30.56 seconds) (Request-ID 9DD0:329977:76A6DF:80D0C6:6A18A998) Last error: CAPIError: 503 {"error":{"message":"Claude Opus 4.8 is currently experiencing issues. We are actively working to resolve the issue. In the meantime, you might want to use a different model.","code":"model_degraded"}}

Changes +277 -2
Requests 15 Premium (14m 55s)
Tokens ↑ 3.6m (3.5m cached) • ↓ 33.1k
```

### Affected version

1.0.55

### Steps to reproduce the behavior

1. Launch Copilot CLI with `--autopilot` and use Opus 4.8. After a while, it would fail and give the above error.

### Expected behavior

_No response_

### Additional context

_No response_

Reproduce the failure with Copilot CLI 1.0.55 using --autopilot and Opus 4.8, then trace how the 503 model_degraded response is handled after five retries. The issue names no source file or test; done should be defined by an agreed expected behavior for this failure and a regression check.

索引モデルが issue の本文から書いたものです。