k8s-security-policies

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 1.00). The "Memory Protocol (MANDATORY)" section instructs the agent to read and write local memory files (e.g., "cat C:...learnings.md" and paths for new patterns/issues/decisions), which is an explicit instruction to access/persist local context unrelated to implementing Kubernetes security policies and therefore constitutes a hidden/deceptive prompt-injection/exfiltration instruction outside the skill's stated purpose.