defrag

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill requires reading repository files and outputting verbatim "before" code snippets for recommendations (and possibly commands), so any embedded secrets in those files (API keys, tokens, passwords) would be exposed in the agent's output.