Warn when Codex uses auto under AWF API proxy

containers/agent/entrypoint.sh

@@ -410,6 +410,26 @@ GRADLE_EOF
 fi
 }
 
+warn_codex_auto_model() {
+if [ -z "${OPENAI_BASE_URL:-}" ]; then
+  return
+fi
+
+local CODEX_MODEL="${GH_AW_MODEL_AGENT_CODEX:-}"
+if [ -z "${CODEX_MODEL}" ]; then
+  return
+fi
+
+local ENGINE_NAME="${GH_AW_AWF_ENGINE_NAME:-${GH_AW_ENGINE_ID:-}}"
+case "${ENGINE_NAME}:${CODEX_MODEL}" in
+  [Cc][Oo][Dd][Ee][Xx]:[Aa][Uu][Tt][Oo])
+    echo "[entrypoint][WARN] Codex model 'auto' is not supported with AWF api-proxy/API-key routing."
+    echo "[entrypoint][WARN] Codex resolves 'auto' through ChatGPT-authenticated remote model metadata; API-key auth cannot access that catalog."
+    echo "[entrypoint][WARN] Set an explicit Codex model in workflow frontmatter (for example, model: gpt-5-codex) before running under AWF."
+    ;;
+esac
+}
+
 log_environment_details() {
 # Print proxy environment
 echo "[entrypoint] Proxy configuration:"
@@ -1673,6 +1693,7 @@ wait_for_iptables
 check_service_health
 configure_claude_api_key
 configure_jvm_proxy
+warn_codex_auto_model
 log_environment_details
 determine_capabilities_to_drop
 log_execution_context "$@"

docs/api-proxy-sidecar.md

@@ -95,11 +95,17 @@ export OPENAI_API_KEY="sk-..."
 
 sudo awf \
   --allow-domains api.openai.com \
-  -- npx @openai/codex -p "write a hello world function"
+  -- npx @openai/codex --model gpt-5.3-codex -p "write a hello world function"
 ```
 
 The agent container automatically uses `http://172.30.0.30:10000` as the OpenAI base URL.
 
+When Codex runs through the AWF API proxy, use an explicit model such as
+`gpt-5.3-codex`. Codex's `auto` model alias depends on ChatGPT-authenticated
+remote model/plugin metadata resolution; API-key based proxy routing cannot use
+that catalog even when `chatgpt.com` is allowed, so `auto` can fail with
+`The requested model is not supported`.
+
 ### Claude Code example
 
 ```bash

docs/troubleshooting.md

@@ -147,6 +147,22 @@
    ```
 4. Review [GitHub Enterprise Configuration](enterprise-configuration.md) for the expected endpoint derivation and allowlist behavior.
 
+### Codex `auto` Model Fails Under AWF
+
+**Problem:** A Codex run inside AWF fails with messages such as:
+- `Unknown model auto is used`
+- `The requested model is not supported`
+- `chatgpt authentication required for remote plugin catalog; api key auth is not supported`
+
+**Cause:** Codex's `auto` model alias relies on ChatGPT-authenticated remote
+model/plugin metadata. AWF's API proxy uses API-key credential injection, so the
+ChatGPT catalog lookup cannot be used even if `chatgpt.com` is on the network
+allowlist.
+
+**Solution:** Set an explicit Codex model instead of `auto`, for example
+`model: gpt-5.3-codex` in workflow frontmatter or `codex exec --model
+gpt-5.3-codex ...` for direct CLI usage.
+
 ## Permission Issues
 
 ### iptables Permission Denied

tests/entrypoint-phase-functions.test.sh

@@ -23,6 +23,7 @@ required_functions=(
   check_service_health
   configure_claude_api_key
   configure_jvm_proxy
+  warn_codex_auto_model
   log_environment_details
   determine_capabilities_to_drop
   log_execution_context
@@ -75,6 +76,7 @@ required_calls=(
   'check_service_health'
   'configure_claude_api_key'
   'configure_jvm_proxy'
+  'warn_codex_auto_model'
   'log_environment_details'
   'determine_capabilities_to_drop'
   'log_execution_context "$@"'
@@ -263,6 +265,59 @@ else
   fail "configure_jvm_proxy() aborts when .m2/.gradle exist on a read-only home"
 fi
 
+run_warn_codex_auto_model_fixture() {
+  local output
+  output="$(
+    env \
+      AWF_API_PROXY_IP="172.30.0.30" \
+      OPENAI_BASE_URL="http://172.30.0.30:10000" \
+      GH_AW_AWF_ENGINE_NAME="codex" \
+      GH_AW_MODEL_AGENT_CODEX="auto" \
+      bash -c '
+        . "$1"
+        warn_codex_auto_model
+      ' _ "${ENTRYPOINT}" 2>&1
+  )"
+  case "${output}" in
+    *"Codex model 'auto' is not supported with AWF api-proxy/API-key routing."*\
+*"Set an explicit Codex model in workflow frontmatter"*)
+      ;;
+    *)
+      return 1
+      ;;
+  esac
+
+  output="$(
+    env \
+      AWF_API_PROXY_IP="172.30.0.30" \
+      GH_AW_AWF_ENGINE_NAME="codex" \
+      GH_AW_MODEL_AGENT_CODEX="gpt-5-codex" \
+      bash -c '
+        . "$1"
+        warn_codex_auto_model
+      ' _ "${ENTRYPOINT}" 2>&1
+  )"
+  [ -z "${output}" ]
+
+  output="$(
+    env \
+      AWF_API_PROXY_IP="172.30.0.30" \
+      GH_AW_AWF_ENGINE_NAME="codex" \
+      GH_AW_MODEL_AGENT_CODEX="auto" \
+      bash -c '
+        . "$1"
+        warn_codex_auto_model
+      ' _ "${ENTRYPOINT}" 2>&1
+  )"
+  [ -z "${output}" ]
+}
+
+if run_warn_codex_auto_model_fixture; then
+  pass "warn_codex_auto_model() warns only for codex model auto under api-proxy"
+else
+  fail "warn_codex_auto_model() did not warn correctly for codex model auto under api-proxy"
+fi
+
 echo ""
 echo "Results: ${PASS} passed, ${FAIL} failed"