securing-agentforce

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly directs the runner and LLM-as-judge to read raw agent responses and write them into judged results and the HTML report (including responses and failures), which means any leaked PII, session IDs, or credentials appearing in those responses would be handled and output verbatim by the LLM.