content-translate-article

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The prompt explicitly requires preserving code blocks, technical identifiers, links, and exact source content verbatim in generated translations, so if the Russian source contains secrets (API keys, tokens, passwords) the LLM would be forced to reproduce them in its output, creating an exfiltration risk.