security-audit

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill requires scanning for hardcoded secrets and producing an evidence-backed report (file paths, line numbers, and an "Evidence Index") but does not instruct redaction, so the LLM may need to include secret values verbatim in its output.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and ingests external, untrusted content as part of its runtime workflow — for example the mandatory tool-installer step and Gemini checks/install (references/tool-installer.md) and the Gemini analysis (references/gemini-analysis.md) which run "gemini prompt" and "gemini extensions install https://github.com/gemini-cli-extensions/security", and the dependency/trivy/gitleaks audit steps (references/dependency-audit.md, references/gitleaks.md) that pull public registry/scan data — all of which are read and integrated into scoring and actions, allowing indirect prompt-injection from third-party sources.