when-validating-code-works-use-functionality-audit

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The content contains high-risk patterns that could enable intentional data exfiltration and supply‑chain abuse (storing raw/base64-encoded logs and artifacts to an external "memory" service, running npx-installed third‑party/alpha packages and agent orchestration), so while not an explicit encoded backdoor, it exposes clear avenues for deliberate malicious behavior if the claude-flow tool or memory backend is untrusted.