health

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill's Step 1 bash explicitly cats settings.local.json, conversation files, and other local files and then requires pasting the full Step 1 output inline into subagent prompts, which will include any API keys/tokens/passwords present verbatim, creating an exfiltration risk.