address-findings

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly requires reading and outputting "full source file contents" and evidence excerpts (which can include hard-coded API keys, tokens, or passwords), so the LLM may be forced to include secret values verbatim in its outputs.