From d94c65bcb214a9fff6d24f0ae29e1de85ce30247 Mon Sep 17 00:00:00 2001
From: Adam Moussa <adam@seahavenind.com>
Date: Wed, 8 Jul 2026 16:30:42 -0400
Subject: [PATCH] fix(scanner): skip gitignored cdk.out synth output in checkov
 scan (INFRA-144)

checkov scanned cdk.out/<stack>.template.json, which is gitignored generated
synth output. A dev with a stale cdk.out lying around would false-block unrelated
pushes on CKV_AWS_111 raised against CDK-generated roles (LogRetention, asset
publishing) that are not authored source. Broaden the checkov --skip-path from
cdk.out/asset. to the whole cdk.out/ tree so it treats synth output the same as
semgrep (--exclude cdk.out) and cfn-lint (cdk.out prune) already do.

Authored IaC checkov parses (SAM/CFN template.yaml, Terraform) is tracked source
and is still fully scanned; verified a planted wildcard IAM policy in tracked
source still trips CKV_AWS_111 and blocks.
---
 review.sh | 17 +++++++++++------
 1 file changed, 11 insertions(+), 6 deletions(-)

diff --git a/review.sh b/review.sh
index 73d294a..2314db5 100755
--- a/review.sh
+++ b/review.sh
@@ -121,12 +121,17 @@ else note_missing gitleaks "brew install gitleaks"; fi
 if command -v checkov >/dev/null; then
   CKALL="$TMP/ck.json"; echo '[]' > "$CKALL"
   for p in $SCOPE_PATHS; do
-    # --skip-path is a regex over the file path. Skip only the cdk.out asset.<hash>/
-    # dependency bundles (the stall cause) + vendored/generated dirs — but KEEP
-    # scanning cdk.out/<stack>.template.json, which is the real deploy artifact
-    # (dropping it would silence genuine S3/IAM IaC findings). asset is anchored to
-    # cdk.out so a source file literally named asset.* is not also excluded.
-    if [ -d "$p" ]; then RAW="$(checkov -d "$p" --skip-path 'cdk\.out/asset\.' --skip-path node_modules --skip-path '\.claude' --skip-path '\.venv' --skip-path venv --skip-path '\.aws-sam' --skip-path dist --skip-path build -o json --compact --quiet 2>/dev/null || true)"
+    # --skip-path is a regex over the file path. Skip the ENTIRE cdk.out/ synth tree
+    # plus other generated/vendored dirs. cdk.out/ is gitignored generated output, not
+    # source — gating pushes on it false-blocks unrelated work (INFRA-144): checkov
+    # raises CKV_AWS_111 (and similar) on CDK-generated roles (LogRetention, asset
+    # publishing, bootstrap) that only exist post-synth and that a dev may have lying
+    # around from a stale `cdk synth`. Authored IaC that checkov actually parses
+    # (SAM/CFN template.yaml, Terraform) is TRACKED source and is still fully scanned;
+    # CDK synth output is covered by the agentic /sh-security-review pass, not this
+    # deterministic gate. Matches semgrep's `--exclude cdk.out` and cfn-lint's cdk.out
+    # prune, so all three scanners now treat synth output consistently.
+    if [ -d "$p" ]; then RAW="$(checkov -d "$p" --skip-path 'cdk\.out/' --skip-path node_modules --skip-path '\.claude' --skip-path '\.venv' --skip-path venv --skip-path '\.aws-sam' --skip-path dist --skip-path build -o json --compact --quiet 2>/dev/null || true)"
     else RAW="$(checkov -f "$p" -o json --compact --quiet 2>/dev/null || true)"; fi
     [ -z "$RAW" ] && continue
     FC="$(echo "$RAW" | jq '[ (if type=="array" then .[] else . end).results.failed_checks // [] ] | add // []' 2>/dev/null || echo '[]')"