humankit

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill explicitly instructs the agent to leave frontmatter, code blocks, tables, and data unchanged and to print or write the rewritten text back, so if the input/file contains API keys, tokens, or passwords those values will be preserved and can be output verbatim — creating an exfiltration risk.