walter-programmatic-seo

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The prompt explicitly requires preserving H1 phrases verbatim from CSV rows and printing row contents in a summary, so if any row contained secrets the model would be instructed to output them unchanged, creating an exfiltration risk.