Skip to content

fix(strix): route openai-direct fallback to the OpenAI API base #1299

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
seonghobae wants to merge 4 commits into from
Closed

fix(strix): route openai-direct fallback to the OpenAI API base #1299

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
5ecbebb
fix(strix): route openai-direct fallback to the OpenAI API base
seonghobae Aug 24, 2026
84eb379
fix(strix): wire direct OpenAI fallback endpoint
seonghobae Aug 24, 2026
0d35300
test(strix): cover direct OpenAI fallback endpoint
seonghobae Aug 24, 2026
b92648e
test(strix): cover GitHub Models OpenAI fallback
seonghobae Aug 24, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 5 additions & 1 deletion .github/workflows/strix.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -674,7 +674,7 @@ jobs:
echo "LLM_API_BASE_FILE=$llm_api_base_file" >> "$GITHUB_ENV"

- name: Prepare GitHub Models fallback credentials
if: steps.gate.outputs.provider_mode == 'openai_direct' || steps.gate.outputs.provider_mode == 'openrouter' || steps.gate.outputs.provider_mode == 'nvidia_nim'
if: steps.gate.outputs.provider_mode == 'github_models' || steps.gate.outputs.provider_mode == 'openai_direct' || steps.gate.outputs.provider_mode == 'openrouter' || steps.gate.outputs.provider_mode == 'nvidia_nim'
env:
GITHUB_MODELS_FALLBACK_TOKEN: ${{ secrets.STRIX_GITHUB_MODELS_TOKEN || github.token }}
OPENAI_FALLBACK_KEY: ${{ secrets.STRIX_OPENAI_API_KEY || secrets.OPENAI_API_KEY }}
Expand Down Expand Up @@ -706,6 +706,9 @@ jobs:
openai_fallback_key_file="$RUNNER_TEMP/openai_fallback_key.txt"
printf '%s' "$openai_trimmed" > "$openai_fallback_key_file"
echo "STRIX_OPENAI_FALLBACK_KEY_FILE=$openai_fallback_key_file" >> "$GITHUB_ENV"
openai_fallback_api_base_file="$RUNNER_TEMP/openai_fallback_api_base.txt"
printf '%s' 'https://api.openai.com/v1' > "$openai_fallback_api_base_file"
echo "STRIX_OPENAI_FALLBACK_API_BASE_FILE=$openai_fallback_api_base_file" >> "$GITHUB_ENV"
fi

- name: Prepare Vertex AI credentials
Expand Down Expand Up @@ -838,6 +841,7 @@ jobs:
STRIX_GITHUB_MODELS_API_BASE_FILE: ${{ env.STRIX_GITHUB_MODELS_API_BASE_FILE }}
STRIX_GITHUB_MODELS_KEY_FILE: ${{ env.STRIX_GITHUB_MODELS_KEY_FILE }}
STRIX_OPENAI_FALLBACK_KEY_FILE: ${{ env.STRIX_OPENAI_FALLBACK_KEY_FILE }}
STRIX_OPENAI_FALLBACK_API_BASE_FILE: ${{ env.STRIX_OPENAI_FALLBACK_API_BASE_FILE }}
STRIX_FAIL_ON_PROVIDER_SIGNAL: "1"
STRIX_VERTEX_FALLBACK_MODELS: ""
NPM_CONFIG_IGNORE_SCRIPTS: "true"
Expand Down
9 changes: 8 additions & 1 deletion scripts/ci/strix_quick_gate.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2417,7 +2417,14 @@ resolved_llm_api_base_for_model() {

local api_base_file="$LLM_API_BASE_FILE"
local api_base_file_name="LLM_API_BASE_FILE"
if is_github_models_model "$model" && [ -n "${STRIX_GITHUB_MODELS_API_BASE_FILE:-}" ]; then
if is_explicit_openai_model "$model" && [ -n "${STRIX_OPENAI_FALLBACK_API_BASE_FILE:-}" ]; then
# Cross-provider fallback: openai-direct/* candidates must reach the
# direct OpenAI API even when the primary provider selected a
# different LLM_API_BASE_FILE endpoint (e.g. NVIDIA NIM). Without
# this the fallback hits the primary gateway and 404s.
api_base_file="$STRIX_OPENAI_FALLBACK_API_BASE_FILE"
api_base_file_name="STRIX_OPENAI_FALLBACK_API_BASE_FILE"
Comment thread
devin-ai-integration[bot] marked this conversation as resolved.
Comment thread
devin-ai-integration[bot] marked this conversation as resolved.
elif is_github_models_model "$model" && [ -n "${STRIX_GITHUB_MODELS_API_BASE_FILE:-}" ]; then
Comment thread
seonghobae marked this conversation as resolved.
# Cross-provider fallback: when the active primary provider uses a
# different API base (for example OpenRouter), github_models/* fallback
# attempts must still route through the GitHub Models inference endpoint.
Comment thread
seonghobae marked this conversation as resolved.
Expand Down
138 changes: 138 additions & 0 deletions scripts/ci/test_strix_quick_gate.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -365,6 +365,9 @@ assert_strix_workflow_pr_trigger_hardened() {
assert_file_contains "$workflow_file" "steps.gate.outputs.provider_mode == 'nvidia_nim' && 'nvidia_nim/nvidia/llama-3.3-nemotron-super-49b-v1.5 openai-direct/gpt-5.6-luna'" "strix workflow gives NVIDIA NIM scans contracted fallbacks"
assert_file_not_contains "$workflow_file" "STRIX_FALLBACK_MODELS: \${{ steps.gate.outputs.provider_mode == 'github_models' && 'github_models/openai/o3" "strix workflow fallback list must not depend on GitHub Models, which is in platform-wide retirement"
assert_file_contains "$workflow_file" "Prepare GitHub Models fallback credentials" "strix workflow provisions GitHub Models fallback credentials for direct-OpenAI scans"
assert_file_contains "$workflow_file" "steps.gate.outputs.provider_mode == 'github_models' || steps.gate.outputs.provider_mode == 'openai_direct'" "strix workflow provisions direct-OpenAI fallback credentials for GitHub Models scans"
assert_file_contains "$workflow_file" "STRIX_OPENAI_FALLBACK_API_BASE_FILE" "strix workflow routes direct-OpenAI fallbacks through a trusted API base file"
assert_file_contains "$workflow_file" "https://api.openai.com/v1" "strix workflow uses the OpenAI platform endpoint for direct fallbacks"
assert_file_contains "$GATE_SCRIPT" "STRIX_GITHUB_MODELS_KEY_FILE" "strix gate reads the optional GitHub Models fallback key file"
assert_file_contains "$GATE_SCRIPT" "STRIX_GITHUB_MODELS_API_BASE_FILE" "strix gate routes github_models fallback models through the GitHub Models endpoint"
assert_file_not_contains "$workflow_file" 'github_models/deepseek/deepseek-r1-0528 | github_models/deepseek/deepseek-v3-0324)' "strix workflow keeps DeepSeek GitHub Models restricted to fallback-only routing"
Expand Down Expand Up @@ -3421,6 +3424,54 @@ REPORT
;;
esac
;;
nvidia-nim-openai-direct-fallback-success)
case "${STRIX_LLM:-}" in
nvidia_nim/primary-model)
if [ "${LLM_API_KEY:-}" != "dummy" ] || [ "${LLM_API_BASE:-}" != "https://integrate.api.nvidia.com/v1" ]; then
echo "unexpected NVIDIA NIM primary credentials or endpoint" >&2
exit 17
fi
echo "nvidia_nim.RateLimitError: Error code: 429 - provider quota exhausted"
exit 1
;;
openai/gpt-5.6-luna)
if [ "${LLM_API_KEY:-}" != "openai-fallback-key" ] || [ "${LLM_API_BASE:-}" != "https://api.openai.com/v1" ]; then
echo "unexpected direct-OpenAI fallback credentials or endpoint" >&2
exit 18
fi
echo "scan ok after direct-OpenAI fallback"
exit 0
;;
*)
echo "unexpected NVIDIA-to-OpenAI fallback model ${STRIX_LLM:-}" >&2
exit 19
;;
esac
;;
github-models-openai-direct-fallback-success)
case "${STRIX_LLM:-}" in
openai/gpt-5)
if [ "${LLM_API_KEY:-}" != "dummy" ] || [ "${LLM_API_BASE:-}" != "https://models.github.ai/inference" ]; then
echo "unexpected GitHub Models primary credentials or endpoint" >&2
exit 20
fi
echo "openai.RateLimitError: Error code: 429 - GitHub Models quota exhausted"
exit 1
;;
openai/gpt-5.6-luna)
if [ "${LLM_API_KEY:-}" != "openai-fallback-key" ] || [ "${LLM_API_BASE:-}" != "https://api.openai.com/v1" ]; then
echo "unexpected GitHub-to-OpenAI fallback credentials or endpoint" >&2
exit 21
fi
echo "scan ok after GitHub-to-OpenAI fallback"
exit 0
;;
*)
echo "unexpected GitHub-to-OpenAI fallback model ${STRIX_LLM:-}" >&2
exit 22
;;
esac
;;
vertex-all-notfound)
echo "Error: litellm.NotFoundError: Vertex_aiException - x"
echo '"status": "NOT_FOUND"'
Expand Down Expand Up @@ -5648,6 +5699,18 @@ PY
env_cmd+=(STRIX_GITHUB_MODELS_API_BASE_FILE="$tmp_dir/github_models_api_base.txt")
env_cmd+=(STRIX_GITHUB_MODELS_KEY_FILE="$tmp_dir/github_models_key.txt")
fi
if [ "$scenario" = "nvidia-nim-openai-direct-fallback-success" ]; then
printf '%s' 'openai-fallback-key' >"$tmp_dir/openai_fallback_key.txt"
printf '%s' 'https://api.openai.com/v1' >"$tmp_dir/openai_fallback_api_base.txt"
env_cmd+=(STRIX_OPENAI_FALLBACK_KEY_FILE="$tmp_dir/openai_fallback_key.txt")
env_cmd+=(STRIX_OPENAI_FALLBACK_API_BASE_FILE="$tmp_dir/openai_fallback_api_base.txt")
fi
if [ "$scenario" = "github-models-openai-direct-fallback-success" ]; then
printf '%s' 'openai-fallback-key' >"$tmp_dir/openai_fallback_key.txt"
printf '%s' 'https://api.openai.com/v1' >"$tmp_dir/openai_fallback_api_base.txt"
env_cmd+=(STRIX_OPENAI_FALLBACK_KEY_FILE="$tmp_dir/openai_fallback_key.txt")
env_cmd+=(STRIX_OPENAI_FALLBACK_API_BASE_FILE="$tmp_dir/openai_fallback_api_base.txt")
fi
if [ "$min_fail_severity" = "__UNSET__" ]; then
local next_env_cmd=()
local env_pair
Expand Down Expand Up @@ -5902,6 +5965,68 @@ run_gate_case_allow_provider_signal() {
run_gate_case_with_provider_signal_mode "0" "$@"
}

run_nvidia_nim_openai_direct_fallback_case() {
run_gate_case_allow_provider_signal "nvidia-nim-openai-direct-fallback-success" \
"nvidia_nim/primary-model" \
"" \
"0" \
"REGEX:Strix quick scan succeeded with fallback model 'openai-direct/gpt-5.6-luna' in [0-9]+s\\." \
"2" \
"nvidia_nim/primary-model|openai/gpt-5.6-luna" \
"https://integrate.api.nvidia.com/v1|https://api.openai.com/v1" \
"nvidia_nim" \
"https://integrate.api.nvidia.com/v1" \
"" \
"0" \
"CRITICAL" \
"0" \
"" \
"" \
"1200" \
"0" \
"" \
"" \
"" \
"" \
"0" \
"" \
"" \
"" \
"__SAME_AS_FALLBACK_MODELS__" \
"openai-direct/gpt-5.6-luna"
}

run_github_models_openai_direct_fallback_case() {
run_gate_case_allow_provider_signal "github-models-openai-direct-fallback-success" \
"openai/gpt-5" \
"" \
"0" \
"REGEX:Strix quick scan succeeded with fallback model 'openai-direct/gpt-5.6-luna' in [0-9]+s\\." \
"2" \
"openai/gpt-5|openai/gpt-5.6-luna" \
"https://models.github.ai/inference|https://api.openai.com/v1" \
"openai" \
"https://models.github.ai/inference" \
"" \
"0" \
"CRITICAL" \
"0" \
"" \
"" \
"1200" \
"0" \
"" \
"" \
"" \
"" \
"0" \
"" \
"" \
"" \
"__SAME_AS_FALLBACK_MODELS__" \
"openai-direct/gpt-5.6-luna"
}

run_github_models_http410_case() {
local scenario="$1"
local expected_exit="$2"
Expand Down Expand Up @@ -6124,6 +6249,12 @@ run_filtered_gate_case_if_requested() {
"" \
"github_models/openai/o3"
;;
nvidia-nim-openai-direct-fallback-success)
run_nvidia_nim_openai_direct_fallback_case
;;
github-models-openai-direct-fallback-success)
run_github_models_openai_direct_fallback_case
;;
gemini-timeout-fallback-success)
run_gate_case_allow_provider_signal "gemini-timeout-fallback-success" \
"gemini/timeout-fallback-primary" \
Expand Down Expand Up @@ -12514,6 +12645,13 @@ run_gate_case "openai-direct-quota-github-models-fallback-success" \
"" \
"github_models/openai/o3"

# NVIDIA NIM primary scans must route an explicit direct-OpenAI fallback to
# the OpenAI endpoint and its own credential, not the primary provider input.
run_nvidia_nim_openai_direct_fallback_case

# GitHub Models primary scans use the same direct-OpenAI fallback contract.
run_github_models_openai_direct_fallback_case

run_gate_case "github-models-fallback-success-deepseek-v3" \
"vertex_ai/missing-primary" \
"github_models/deepseek/deepseek-r1-0528 github_models/deepseek/deepseek-v3-0324" \
Expand Down
Loading