code-simplifier

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt requires preserving exact functionality when autonomously modifying recently changed code, so if that code contains hard-coded API keys/passwords the agent would need to read and reproduce those secret values verbatim in its output, creating an exfiltration risk.