source-code-scanning

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt explicitly instructs scanning for hardcoded secrets and to include "evidence" and raw tool output (e.g., trufflehog/gitleaks JSON, PoC) which would contain secret values verbatim, so the LLM would be expected to handle and likely output those secrets.

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

• Third-party content exposure detected (high risk: 0.80). The skill's required workflow and referenced docs explicitly instruct fetching and checking public, user-generated sources (e.g., running "trufflehog github --repo https://github.com/org/repo" in reference/secrets-detection.md and manual CVE verification via NVD/OSV URLs in reference/dependency-cve-scanning.md), which the agent would read and act on to influence vulnerability triage and remediation decisions.