test(cli): parse Click JSON output robustly in tests

tests/cli_output.py

@@ -0,0 +1,37 @@
+"""Helpers for asserting JSON output from Click CLI tests."""
+
+from __future__ import annotations
+
+import json
+from typing import Any
+
+
+def parse_click_json_output(output: str) -> dict[str, Any]:
+    """Parse JSON from `CliRunner` output, tolerating prefixed stderr log lines.
+
+    Click's test runner can merge stderr into `result.output`, so scans that emit
+    findings may prepend log lines before the JSON payload even though a real
+    subprocess keeps stdout and stderr separate.
+    """
+    try:
+        parsed = json.loads(output)
+    except json.JSONDecodeError:
+        decoder = json.JSONDecoder()
+        for index, char in enumerate(output):
+            if char != "{":
+                continue
+
+            try:
+                parsed, end = decoder.raw_decode(output, index)
+            except json.JSONDecodeError:
+                continue
+
+            if output[end:].strip():
+                continue
+            break
+        else:
+            raise
+
+    if not isinstance(parsed, dict):
+        raise TypeError(f"Expected a JSON object, got {type(parsed).__name__}")
+    return parsed

tests/conftest.py

@@ -116,6 +116,7 @@ def pytest_runtest_setup(item):
             "test_sevenzip_scanner.py",  # 7-Zip archive scanner tests
             "test_regression_corpus.py",  # malicious/safe fixture regression gate
             "test_nested_pickle_integration.py",  # nested pickle false-positive/true-positive integration tests
+            "test_cli_output.py",  # CliRunner JSON parsing helper regression tests
         ]
 
         # Check if this is an allowed test file

tests/test_cli.py

@@ -1,6 +1,8 @@
 import json
 import os
 import re
+import subprocess
+import sys
 from pathlib import Path
 from typing import Any
 from unittest.mock import patch
@@ -13,6 +15,7 @@
 from modelaudit.cli import cli, expand_paths, format_text_output
 from modelaudit.core import scan_model_directory_or_file
 from modelaudit.models import ModelAuditResultModel, create_initial_audit_result
+from tests.cli_output import parse_click_json_output
 
 
 def strip_ansi(text: str) -> str:
@@ -163,18 +166,42 @@ def test_scan_does_not_auto_load_untrusted_local_config(tmp_path: Path) -> None:
 
     model_file = tmp_path / "evil.tar"
     with tarfile.open(model_file, "w") as tar:
-        payload = tmp_path / "payload.txt"
-        payload.write_text("content")
-        tar.add(payload, arcname="../evil.txt")
+        payload_file = tmp_path / "payload.txt"
+        payload_file.write_text("content")
+        tar.add(payload_file, arcname="../evil.txt")
 
     (tmp_path / ".modelaudit.toml").write_text('suppress = ["S405"]\n')
 
     runner = CliRunner()
     result = runner.invoke(cli, ["scan", str(model_file), "--format", "json"], catch_exceptions=False)
 
     assert result.exit_code == 1
-    payload = json.loads(result.output)
-    assert any(issue.get("rule_code") == "S405" for issue in payload.get("issues", []))
+    output_payload = parse_click_json_output(result.output)
+    assert any(issue.get("rule_code") == "S405" for issue in output_payload.get("issues", []))
+
+
+def test_scan_json_subprocess_separates_logs_from_stdout_for_findings(tmp_path: Path) -> None:
+    """Real process execution should keep JSON stdout parseable even when findings are logged."""
+    import tarfile
+
+    model_file = tmp_path / "evil.tar"
+    with tarfile.open(model_file, "w") as tar:
+        payload_file = tmp_path / "payload.txt"
+        payload_file.write_text("content")
+        tar.add(payload_file, arcname="../evil.txt")
+
+    completed = subprocess.run(
+        [sys.executable, "-m", "modelaudit", "scan", str(model_file), "--format", "json"],
+        check=False,
+        capture_output=True,
+        text=True,
+    )
+
+    assert completed.returncode == 1
+    assert completed.stdout.lstrip().startswith("{")
+    output_payload = json.loads(completed.stdout)
+    assert any(issue.get("rule_code") == "S405" for issue in output_payload.get("issues", []))
+    assert "[S405]" in completed.stderr
 
 
 def test_scan_can_apply_local_config_once_when_confirmed(tmp_path: Path, monkeypatch: pytest.MonkeyPatch) -> None:
@@ -409,7 +436,7 @@ def test_scan_json_output(tmp_path):
     # For JSON output, we should be able to parse the output as JSON
     # regardless of the exit code
     try:
-        output_json = json.loads(result.output)
+        output_json = parse_click_json_output(result.output)
         assert "files_scanned" in output_json
         assert "issues" in output_json
         assert output_json["files_scanned"] == 1
@@ -469,7 +496,7 @@ def test_scan_json_to_stdout_no_progress_interference(tmp_path):
 
     # Output should be valid JSON when going to stdout (no progress interference)
     try:
-        output_json = json.loads(result.output)
+        output_json = parse_click_json_output(result.output)
         assert "files_scanned" in output_json
         assert "issues" in output_json
     except json.JSONDecodeError:
@@ -932,7 +959,7 @@ def file_generator():
 
     # Verify content_hash is in JSON output
     try:
-        output_json = json.loads(result.output)
+        output_json = parse_click_json_output(result.output)
         assert "content_hash" in output_json
         assert output_json["content_hash"] == "a" * 64
         assert output_json["files_scanned"] == 3

tests/test_cli_output.py

@@ -0,0 +1,25 @@
+from __future__ import annotations
+
+import json
+
+import pytest
+
+from tests.cli_output import parse_click_json_output
+
+
+def test_parse_click_json_output_handles_prefixed_log_lines() -> None:
+    """Click tests may prefix stderr logs before the JSON body."""
+    payload = {"files_scanned": 1, "issues": [{"rule_code": "S405"}]}
+    output = (
+        "2026-03-20 16:16:35,821 - modelaudit.scanners - CRITICAL - "
+        "[S405] [CRITICAL] (evil.tar:../evil.txt): Archive entry ../evil.txt attempted path traversal\n"
+        f"{json.dumps(payload, indent=2)}\n"
+    )
+
+    assert parse_click_json_output(output) == payload
+
+
+def test_parse_click_json_output_rejects_non_object_json() -> None:
+    """CLI scan JSON should be an object payload, not a scalar or array."""
+    with pytest.raises(TypeError, match="Expected a JSON object"):
+        parse_click_json_output("[]")

tests/test_nested_pickle_integration.py

@@ -5,7 +5,6 @@
 capabilities, testing both malicious detection and false positive prevention.
 """
 
-import json
 import os
 import tempfile
 from pathlib import Path
@@ -17,6 +16,7 @@
 from modelaudit.core import determine_exit_code, scan_model_directory_or_file
 from modelaudit.scanners.base import IssueSeverity
 from modelaudit.scanners.pickle_scanner import PickleScanner
+from tests.cli_output import parse_click_json_output
 
 
 class TestNestedPickleIntegration:
@@ -197,7 +197,7 @@ def test_cli_nested_pickle_detection(self, pickles_dir):
         json_result = runner.invoke(cli, ["scan", str(malicious_file), "--format", "json"])
         assert json_result.exit_code == 1, f"JSON format should detect malicious file. Output: {json_result.output}"
 
-        output_data = json.loads(json_result.output)
+        output_data = parse_click_json_output(json_result.output)
         nested_issues = [
             issue
             for issue in output_data["issues"]