security-scanner

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly scans config files like .env and requires including "the actual vulnerable code snippet as evidence" and flagging hardcoded secrets, which forces the agent to read and potentially output secret values verbatim in its report.

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

• Third-party content exposure detected (medium risk: 0.30). Skill.md Step 1/2/3 ingest outsider-provided code by cloning a user-supplied GitHub repository URL and then reading/grepping files to produce findings, so a malicious repo can include free text/code that the runtime LLM processes.