writing-clearly-and-concisely

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill mandates preserving code, commands, paths, and verbatim load-bearing tokens and returning only the revised text, so if the user input includes secrets (API keys, tokens, passwords) the model is effectively required to output them verbatim, creating a high exfiltration risk.