feat: Enhance ephemeral review process with reasoning effort handling and retry logic

src/gitea_codex_bot/workers/container_runner.py

@@ -31,20 +31,25 @@ def run_review_ephemeral(
     gitea = GiteaClient(settings)
     prompt, _diff_context, repo_cfg = prepare_review_prompt(settings, gitea, repo, pr_number, command)
     container_name = f"codex-review-{uuid.uuid4().hex[:12]}"
-    install_and_run = _build_install_and_run_command(settings)
     extra_env: dict[str, str] = {}
     if settings.codex_auth_mode == "chatgpt":
         extra_env["CODEX_AUTH_JSON_B64"] = _load_codex_auth_json_b64(settings)
-    cmd = _build_docker_command(settings, container_name=container_name, install_and_run=install_and_run)
     try:
-        completed = subprocess.run(
-            cmd,
-            input=prompt,
-            text=True,
-            check=False,
-            capture_output=True,
-            timeout=settings.max_review_minutes * 60,
-            env={**os.environ, **extra_env},
+        completed = _run_ephemeral_container(
+            settings,
+            container_name=container_name,
+            prompt=prompt,
+            extra_env=extra_env,
+            include_reasoning_effort=True,
+        )
+        if _needs_reasoning_effort_compat_retry(completed):
+            logger.info("Ephemeral runner does not support --reasoning-effort; retrying without it.")
+            completed = _run_ephemeral_container(
+                settings,
+                container_name=container_name,
+                prompt=prompt,
+                extra_env=extra_env,
+                include_reasoning_effort=False,
             )
         if completed.returncode != 0:
             raise RuntimeError(_format_runner_failure(completed))
@@ -56,7 +61,28 @@ def run_review_ephemeral(
         return _ephemeral_runner_failure_result(exc, settings.codex_auth_mode), repo_cfg
 
 
-def _build_install_and_run_command(settings: Settings) -> str:
+def _run_ephemeral_container(
+    settings: Settings,
+    *,
+    container_name: str,
+    prompt: str,
+    extra_env: dict[str, str],
+    include_reasoning_effort: bool,
+) -> subprocess.CompletedProcess[str]:
+    install_and_run = _build_install_and_run_command(settings, include_reasoning_effort=include_reasoning_effort)
+    cmd = _build_docker_command(settings, container_name=container_name, install_and_run=install_and_run)
+    return subprocess.run(
+        cmd,
+        input=prompt,
+        text=True,
+        check=False,
+        capture_output=True,
+        timeout=settings.max_review_minutes * 60,
+        env={**os.environ, **extra_env},
+    )
+
+
+def _build_install_and_run_command(settings: Settings, *, include_reasoning_effort: bool = True) -> str:
     steps = ["set -euo pipefail"]
     if settings.codex_auth_mode == "chatgpt":
         steps.extend(
@@ -77,12 +103,19 @@ def _build_install_and_run_command(settings: Settings) -> str:
     codex_exec_parts = ["codex exec --skip-git-repo-check --json"]
     if model:
         codex_exec_parts.append(f"-m {shlex.quote(model)}")
-    if reasoning_effort:
+    if include_reasoning_effort and reasoning_effort:
         codex_exec_parts.append(f"--reasoning-effort {shlex.quote(reasoning_effort)}")
     steps.append(" ".join(codex_exec_parts))
     return "; ".join(steps)
 
 
+def _needs_reasoning_effort_compat_retry(completed: subprocess.CompletedProcess[str]) -> bool:
+    if completed.returncode == 0:
+        return False
+    stderr_text = completed.stderr or ""
+    return "unexpected argument '--reasoning-effort' found" in stderr_text
+
+
 def _build_docker_command(settings: Settings, *, container_name: str, install_and_run: str) -> list[str]:
     cmd = [
         "docker",

tests/test_container_runner.py

@@ -72,6 +72,14 @@ def test_build_install_command_includes_configured_reasoning_effort(monkeypatch:
     assert "--reasoning-effort medium" in command
 
 
+def test_build_install_command_can_disable_reasoning_effort_flag() -> None:
+    settings = get_settings()
+
+    command = _build_install_and_run_command(settings, include_reasoning_effort=False)
+
+    assert "--reasoning-effort" not in command
+
+
 def test_chatgpt_mode_requires_existing_auth_json(monkeypatch: pytest.MonkeyPatch, tmp_path: Path) -> None:
     missing = tmp_path / "missing-auth.json"
     monkeypatch.setenv("CODEX_AUTH_MODE", "chatgpt")
@@ -157,6 +165,59 @@ def test_run_review_ephemeral_api_key_mode_does_not_fallback_to_host(monkeypatch
     assert "API-key auth runner failed" in result["summary"]
 
 
+def test_run_review_ephemeral_retries_without_reasoning_effort_when_unsupported(monkeypatch: pytest.MonkeyPatch) -> None:
+    get_settings.cache_clear()
+    settings = get_settings()
+
+    monkeypatch.setattr(
+        "gitea_codex_bot.workers.container_runner.prepare_review_prompt",
+        lambda *_args, **_kwargs: ("prompt", {"diff": ""}, object()),
+    )
+    monkeypatch.setattr("gitea_codex_bot.workers.container_runner.GiteaClient", lambda _settings: object())
+
+    calls: list[list[str]] = []
+
+    def _fake_run(cmd, *args, **kwargs):
+        calls.append(cmd)
+        if len(calls) == 1:
+            return type(
+                "Completed",
+                (),
+                {
+                    "returncode": 2,
+                    "stdout": "",
+                    "stderr": "error: unexpected argument '--reasoning-effort' found",
+                },
+            )()
+        return type(
+            "Completed",
+            (),
+            {
+                "returncode": 0,
+                "stdout": '{"verdict":"correct","confidence":0.9,"summary":"ok","findings":[]}\n',
+                "stderr": "",
+            },
+        )()
+
+    monkeypatch.setattr("gitea_codex_bot.workers.container_runner.subprocess.run", _fake_run)
+
+    from gitea_codex_bot.types import ParsedCommand
+
+    result, _repo_cfg = run_review_ephemeral(
+        settings,
+        repo="acme/repo",
+        pr_number=1,
+        command=ParsedCommand(name="review", raw="@codex review"),
+    )
+
+    assert result["verdict"] == "correct"
+    assert len(calls) == 2
+    first_shell = calls[0][-1]
+    second_shell = calls[1][-1]
+    assert "--reasoning-effort" in first_shell
+    assert "--reasoning-effort" not in second_shell
+
+
 def test_parse_codex_exec_stdout_from_stream_item_text_json() -> None:
     stdout = '\n'.join(
         [