building-automated-malware-submission-pipeline

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly ingests public third‑party data from VirusTotal and MalwareBazaar (see SKILL.md Step 2 "Pre‑Screen with Hash Lookups" and scripts/agent.py functions check_virustotal/check_malwarebazaar), and it parses those untrusted responses to decide sandboxing, verdicts, and blocking actions—allowing external content to materially influence tool use and next steps.