writing-upstream-bug-reports

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The prompt mandates copy-pasting "exact commands, exact output" and including full logs/configs/attachments in the report, which can contain API keys, tokens, cookies or passwords and thus would require the LLM to reproduce secret values verbatim.