fix(security): Strix provider 장애를 fail-closed로 유지

.github/workflows/strix.yml

@@ -752,50 +752,20 @@ jobs:
           export "STRIX_PROCESS_${budget_suffix}_SECONDS=$process_budget_seconds"
           export "STRIX_TOTAL_${budget_suffix}_SECONDS=5700"
 
-          # Capture the gate exit code plus its console output. The gate returns
-          # exit 1 both for genuine blocking vulnerabilities AND for
-          # LLM-backend-unavailable outcomes (GitHub Models "Too many requests"
-          # rate limits, OpenAI quota starvation, 413 tokens_limit_reached
-          # token-cap, connection/warm-up failures) that could not complete a scan. A backend outage is CI
-          # infrastructure noise, not a security finding, so it must not fail
-          # the required check and block merges.
+          # Capture the gate exit code plus its console output so the always-run
+          # artifact steps can preserve exact provider and finding evidence.
+          # A provider outage without a vulnerability report is incomplete
+          # security evidence, so the gate's nonzero result must remain nonzero.
           strix_run_log="$RUNNER_TEMP/strix_gate_console.log"
           strix_rc=0
           set +e
           bash "$TRUSTED_STRIX_GATE" 2>&1 | tee "$strix_run_log"
           strix_rc="${PIPESTATUS[0]}"
           set -e
 
-          if [ "$strix_rc" -eq 0 ]; then
-            exit 0
-          fi
-
-          # Preserve configuration failures (exit 2) and any unexpected exit
-          # code as hard failures — only the scan-failure code (1) can be an
-          # infrastructure/backend-unavailability outcome.
-          if [ "$strix_rc" -ne 1 ]; then
-            exit "$strix_rc"
+          if [ "$strix_rc" -ne 0 ]; then
+            echo "::error title=Strix security gate failed::Strix returned exit ${strix_rc}. Provider outages without a vulnerability report are incomplete security evidence and are not converted to success. See the strix-reports artifact and run log."
           fi
-
-          # Recognized signals that the LLM backend was unavailable / starved.
-          backend_unavailable_signal='RateLimitError|Too many requests\. For more on scraping GitHub|exceeded your current quota|insufficient_quota|billing details|"status"[[:space:]]*:[[:space:]]*"RESOURCE_EXHAUSTED"|tokens_limit_reached|Request body too large|Max size:[[:space:]]*[0-9]+[[:space:]]+tokens|Error code:[[:space:]]*413|LLM CONNECTION FAILED|Could not establish connection to the language model|LLM warm-up failed|Configured model and fallback models were unavailable|Configured Vertex model and fallback models were unavailable|emitted provider infrastructure or failure-signal output|before provider infrastructure failure'
-          # Any evidence that a vulnerability was actually reported. Its presence
-          # forces a hard failure so real findings are NEVER downgraded. Keep the
-          # severity branch anchored away from identifiers so environment lines
-          # such as STRIX_FAIL_ON_MIN_SEVERITY do not look like findings.
-          reported_vulnerability_signal='Vulnerabilities[[:space:]]+[1-9]|(^|[^A-Za-z0-9_])severity[[:space:]]*:'
-
-          # Neutral skip only when ALL hold: a backend-unavailability signal is
-          # present and no vulnerability was reported anywhere. This preserves
-          # real security gating while keeping uncontrollable provider outages
-          # from blocking current-head merge progress.
-          if grep -Eiq "$backend_unavailable_signal" "$strix_run_log" \
-            && ! grep -Eiq "$reported_vulnerability_signal" "$strix_run_log"; then
-            echo "::warning title=Strix backend unavailable::Strix could not complete because its LLM backend was unavailable (rate limit / token cap / connection or warm-up failure) before producing a vulnerability report. Treating as a neutral skip so an infrastructure outage does not block merges; genuine findings still fail the check. See the strix-reports artifact and the run log."
-            exit 0
-          fi
-
-          echo "Strix reported security findings or failed for a non-backend reason; failing the required check (gate exit ${strix_rc})." >&2
           exit "$strix_rc"
 
       - name: Collect Strix reports for artifact upload

scripts/ci/test_strix_quick_gate.sh

@@ -284,6 +284,9 @@ assert_strix_workflow_pr_trigger_hardened() {
 	assert_file_contains "$workflow_file" "vertex_ai/gemini-3.1-pro-preview-customtools | vertex_ai/gemini-2.5-flash)" "strix workflow accepts only exact approved organization Vertex AI models"
 	assert_file_contains "$workflow_file" 'STRIX_VERTEX_FALLBACK_MODELS: ""' "strix workflow disables silent Vertex fallbacks so timeout-class failures fail closed"
 	assert_file_contains "$workflow_file" 'STRIX_FAIL_ON_PROVIDER_SIGNAL: "1"' "strix workflow fails closed on timeout, fatal, warning, denied, or provider failure signals"
+	assert_file_contains "$workflow_file" 'Provider outages without a vulnerability report are incomplete security evidence and are not converted to success' "strix workflow keeps provider exhaustion fail-closed"
+	assert_file_not_contains "$workflow_file" "Treating as a neutral skip" "strix workflow must not convert provider exhaustion into a successful security check"
+	assert_file_not_contains "$workflow_file" "backend_unavailable_signal=" "strix workflow must not recognize provider errors in order to downgrade the gate exit code"
 	assert_file_contains "$workflow_file" 'NPM_CONFIG_IGNORE_SCRIPTS: "true"' "strix workflow disables npm lifecycle scripts for untrusted PR scan data"
 	assert_file_contains "$workflow_file" 'PNPM_CONFIG_IGNORE_SCRIPTS: "true"' "strix workflow disables pnpm lifecycle scripts for untrusted PR scan data"
 	assert_file_contains "$workflow_file" 'YARN_ENABLE_SCRIPTS: "false"' "strix workflow disables yarn lifecycle scripts for untrusted PR scan data"
@@ -673,7 +676,7 @@ assert_opencode_review_uses_codegraph_and_gpt5_fallback() {
 	assert_file_contains "$REPO_ROOT/scripts/ci/run_opencode_review_model_pool.sh" "not a generic model-exhaustion message" "opencode review tells models to return concrete missing-evidence findings instead of progress-only output"
 	assert_file_contains "$REPO_ROOT/scripts/ci/run_opencode_review_model_pool.sh" "tokens_limit_reached" "opencode review detects provider context-window overflow"
 	assert_file_contains "$REPO_ROOT/scripts/ci/run_opencode_review_model_pool.sh" "skipping remaining attempts for this model" "opencode review skips same-model retries after context-window overflow"
-	assert_file_contains "$REPO_ROOT/.github/workflows/strix.yml" "exceeded your current quota" "strix wrapper neutralizes quota-only provider failures without vulnerability reports"
+	assert_file_contains "$REPO_ROOT/.github/workflows/strix.yml" 'exit "$strix_rc"' "strix wrapper propagates provider-failure and finding exit codes without downgrading them"
 	assert_file_contains "$REPO_ROOT/scripts/ci/strix_quick_gate.sh" "billing details" "strix quick gate classifies provider quota starvation as infrastructure"
 	assert_file_contains "$workflow_file" 'timeout-minutes: 300' "opencode review target contains evidence, the bounded long-review pool, publication, and cleanup overhead"
 	assert_file_contains "$workflow_file" 'timeout-minutes: 12' "opencode evidence preparation fails closed before it ties up the review queue"

tests/test_required_workflow_queue_contract.py

@@ -1,4 +1,5 @@
 import json
+import os
 import shlex
 import shutil
 import subprocess
@@ -885,20 +886,46 @@ def test_optional_strix_workflow_absence_is_logged_without_failing_lookup() -> N
     assert 'if target_workflow_available "strix.yml"; then' in failed_check_evidence
 
 
-def test_strix_provider_outage_without_findings_is_neutralized() -> None:
+def test_strix_provider_outage_without_findings_fails_closed(tmp_path: Path) -> None:
     workflow = workflow_text("strix.yml")
+    run_step = workflow_step(workflow, "Run Strix (quick)")
+    run_marker = "        run: |\n"
+    run_body = run_step.split(run_marker, 1)[1]
+    script = textwrap.dedent(
+        "\n".join(line[10:] for line in run_body.splitlines())
+    )
 
-    assert "RateLimitError|Too many requests" in workflow
-    assert "exceeded your current quota" in workflow
-    assert "billing details" in workflow
-    assert "LLM warm-up failed" in workflow
-    assert "zero_vulnerabilities_signal" not in workflow
-    assert "(^|[^A-Za-z0-9_])severity[[:space:]]*:" in workflow
-    assert "STRIX_FAIL_ON_MIN_SEVERITY: MEDIUM" in workflow
-    assert "before producing a vulnerability report" in workflow
-    assert "genuine findings still fail the check" in workflow
-    assert (
-        '&& ! grep -Eiq "$reported_vulnerability_signal" "$strix_run_log"' in workflow
+    fake_gate = tmp_path / "fake-strix-gate.sh"
+    fake_gate.write_text(
+        "#!/usr/bin/env bash\n"
+        "echo 'openai.RateLimitError: exceeded your current quota'\n"
+        "echo 'Configured model and fallback models were unavailable.'\n"
+        "exit 1\n",
+        encoding="utf-8",
+    )
+    fake_gate.chmod(0o700)
+
+    result = subprocess.run(
+        ["bash"],
+        input=script,
+        capture_output=True,
+        text=True,
+        env={
+            **os.environ,
+            "RUNNER_TEMP": str(tmp_path),
+            "TRUSTED_STRIX_GATE": str(fake_gate),
+        },
+        check=False,
+    )
+
+    assert result.returncode == 1
+    assert "Provider outages without a vulnerability report" in result.stdout
+    assert "not converted to success" in result.stdout
+    assert "Treating as a neutral skip" not in workflow
+    assert "backend_unavailable_signal=" not in workflow
+    assert (tmp_path / "strix_gate_console.log").read_text(encoding="utf-8") == (
+        "openai.RateLimitError: exceeded your current quota\n"
+        "Configured model and fallback models were unavailable.\n"
     )