python-debugging

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill instructs the agent to inspect/evaluate variables and return their values in JSON (e.g., locals/inspect/eval), which can cause secrets (API keys, tokens, passwords) present in program state to be output verbatim and exfiltrated.