synthesis-response-merger

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt mandates preserving every detail from input responses and explicitly forbids omitting or shortening content, so any API keys, tokens, or passwords present in the inputs would be reproduced verbatim in the merged output, creating a direct secret-exfiltration risk.