fix: catch FileNotFoundError in F1 JSON scorer verifiers

When the agent fails to produce the expected output file (chain.json,
callers.json, etc.), the Python verifier in test.sh crashed with an
unhandled FileNotFoundError. This caused the entire bash script to exit
(due to set -e) without writing the reward file, leading Harbor to throw
RewardFileNotFoundError instead of scoring 0.0.

Added FileNotFoundError and OSError to the except clause in all 16
affected test.sh files (5 design, 8 code-review, 2 build, 1 template).

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: sjarmak

benchmarks/ccb_build/envoy-grpc-server-impl-001/tests/test.sh

@@ -119,8 +119,8 @@ try:
     if not isinstance(reported, list):
         print("Agent output is not a JSON array — scoring as empty.")
         reported = []
-except (json.JSONDecodeError, ValueError) as e:
-    print(f"Malformed JSON in agent output: {e}")
+except (json.JSONDecodeError, ValueError, FileNotFoundError, OSError) as e:
+    print(f"Could not read/parse agent output: {e}")
     reported = []
 
 num_reported = len(reported)

benchmarks/ccb_build/k8s-runtime-object-impl-001/tests/test.sh

@@ -119,8 +119,8 @@ try:
     if not isinstance(reported, list):
         print("Agent output is not a JSON array — scoring as empty.")
         reported = []
-except (json.JSONDecodeError, ValueError) as e:
-    print(f"Malformed JSON in agent output: {e}")
+except (json.JSONDecodeError, ValueError, FileNotFoundError, OSError) as e:
+    print(f"Could not read/parse agent output: {e}")
     reported = []
 
 num_reported = len(reported)

benchmarks/ccb_design/envoy-routeconfig-dep-chain-001/tests/test.sh

@@ -95,8 +95,8 @@ try:
     if not isinstance(reported_steps, list):
         print("Agent output is not a JSON array — scoring as empty.")
         reported_steps = []
-except (json.JSONDecodeError, ValueError) as e:
-    print(f"Malformed JSON in agent output: {e}")
+except (json.JSONDecodeError, ValueError, FileNotFoundError, OSError) as e:
+    print(f"Could not read/parse agent output: {e}")
     reported_steps = []
 
 num_reported = len(reported_steps)

benchmarks/ccb_design/envoy-stream-aggregated-sym-001/tests/test.sh

@@ -84,8 +84,8 @@ try:
     if not isinstance(reported, list):
         print("Agent output is not a JSON array — scoring as empty.")
         reported = []
-except (json.JSONDecodeError, ValueError) as e:
-    print(f"Malformed JSON in agent output: {e}")
+except (json.JSONDecodeError, ValueError, FileNotFoundError, OSError) as e:
+    print(f"Could not read/parse agent output: {e}")
     reported = []
 
 num_reported = len(reported)

benchmarks/ccb_design/k8s-sharedinformer-sym-001/tests/test.sh

@@ -85,8 +85,8 @@ try:
     if not isinstance(reported, list):
         print("Agent output is not a JSON array — scoring as empty.")
         reported = []
-except (json.JSONDecodeError, ValueError) as e:
-    print(f"Malformed JSON in agent output: {e}")
+except (json.JSONDecodeError, ValueError, FileNotFoundError, OSError) as e:
+    print(f"Could not read/parse agent output: {e}")
     reported = []
 
 num_reported = len(reported)

benchmarks/ccb_design/k8s-typemeta-dep-chain-001/tests/test.sh

@@ -95,8 +95,8 @@ try:
     if not isinstance(reported_steps, list):
         print("Agent output is not a JSON array — scoring as empty.")
         reported_steps = []
-except (json.JSONDecodeError, ValueError) as e:
-    print(f"Malformed JSON in agent output: {e}")
+except (json.JSONDecodeError, ValueError, FileNotFoundError, OSError) as e:
+    print(f"Could not read/parse agent output: {e}")
     reported_steps = []
 
 num_reported = len(reported_steps)

benchmarks/ccb_design/terraform-provider-iface-sym-001/tests/test.sh

@@ -85,8 +85,8 @@ try:
     if not isinstance(reported, list):
         print("Agent output is not a JSON array — scoring as empty.")
         reported = []
-except (json.JSONDecodeError, ValueError) as e:
-    print(f"Malformed JSON in agent output: {e}")
+except (json.JSONDecodeError, ValueError, FileNotFoundError, OSError) as e:
+    print(f"Could not read/parse agent output: {e}")
     reported = []
 
 num_reported = len(reported)

benchmarks/ccb_test/aspnetcore-code-review-001/tests/test.sh

@@ -82,7 +82,7 @@ if os.path.isfile(review_path):
         reported = json.loads(raw)
         if not isinstance(reported, list):
             reported = []
-    except (json.JSONDecodeError, ValueError):
+    except (json.JSONDecodeError, ValueError, FileNotFoundError, OSError):
         reported = []
 
 # Build set of expected file paths for matching

benchmarks/ccb_test/calcom-code-review-001/tests/test.sh

@@ -82,7 +82,7 @@ if os.path.isfile(review_path):
         reported = json.loads(raw)
         if not isinstance(reported, list):
             reported = []
-    except (json.JSONDecodeError, ValueError):
+    except (json.JSONDecodeError, ValueError, FileNotFoundError, OSError):
         reported = []
 
 # Build set of expected file paths for matching

benchmarks/ccb_test/curl-security-review-001/tests/test.sh

@@ -89,7 +89,7 @@ if os.path.isfile(review_path):
         reported = json.loads(raw)
         if not isinstance(reported, list):
             reported = []
-    except (json.JSONDecodeError, ValueError):
+    except (json.JSONDecodeError, ValueError, FileNotFoundError, OSError):
         reported = []
 
 # Build set of expected file paths for matching