lgnh-commit-detail

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill mandates staging and committing all working-tree changes and building commit messages from context that includes git diffs, which can force the model to read and reproduce secret values verbatim (including placing them into the heredoc commit message), creating an exfiltration risk.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The document explicitly instructs staging and committing all working-tree changes (including unintended edits) without prompting, which enables deliberate or accidental exfiltration of sensitive data and credential leakage.