feature-implementation

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill mandates reading and reproducing actual code and file contents (including full metadata and "actual code" evidence) and to insert those into analysis outputs and generated plans without any redaction rules, which would force the LLM to include any secrets found in the repo verbatim if present.