analysis

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill's queries explicitly request dumping string contents from binaries (e.g., SELECT content FROM strings, string LIKE '%password%'), which can contain API keys/tokens/passwords and thus would require the model to output secret values verbatim.