Suggested test: codex/openai-model-prefix-stripping

## Motivation

Link to the gh-aw PR: github/gh-aw#60423 — `codexModelID` stripped only the `copilot/` prefix from model identifiers, so `engine.model: openai/gpt-5.3-codex` was forwarded verbatim to the Codex CLI and failed with `model_not_supported_error`. The fix strips `openai/` (case-insensitively) too.

## Proposed test

- **Workflow file**: `test-codex-openai-model-prefix.md`
- **Trigger**: `workflow_dispatch`
- **Engine**: codex (engine-specific fix, not portable to copilot/claude)
- **Safe output**: `create-issue`
- **Variant**: standard

## Minimal test prompt sketch

Configure `engine: { id: codex, model: openai/gpt-5.3-codex }` and have the agent create a simple issue with a fixed title/body. Since the compiler bug affected the model ID actually forwarded to the Codex CLI (not just the compiled lock file), this scenario is only meaningfully exercised when the live engine runs (`--use-samples` would bypass the invocation). Flag this test for inclusion in `main-live` / non-samples runs only.

## New fixtures or secrets needed

None — uses existing repo and codex engine credentials already configured for other `test-codex-*` workflows.

## Notes

No existing `test-codex-*` workflow sets an explicit `model:` with a provider prefix. Open question: should this run under `--use-samples` at all, or be excluded (like `test-copilot-custom-safe-outputs`) since its whole value is exercising real Codex CLI invocation with the corrected model string?

> Generated by [🔍 Suggest New E2E Tests](https://github.com/githubnext/gh-aw-test/actions/runs/34747998916) · copilot · auto · 71.7 AIC · ⌖ 5.41 AIC · ⊞ 8.4K · [◷](https://github.com/search?q=repo%3Agithubnext%2Fgh-aw-test+is%3Aissue+%22gh-aw-workflow-call-id%3A+githubnext%2Fgh-aw-test%2Fsuggest-new-e2e-tests%22&type=issues)

Create test-codex-openai-model-prefix-stripping.md, using the existing test-codex-* workflows as a starting point. Configure the codex engine with model openai/gpt-5.3-codex and a fixed create-issue output, then run it through main-live or another non-samples path with the existing Codex credentials. Done means the live Codex invocation succeeds and creates the expected issue.

Written by the indexing model from the issue text.