analyzing-ransomware-leak-site-intelligence

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches public, untrusted leak-site data (e.g., the Ransomwatch raw GitHub URL in SKILL.md: https://raw.githubusercontent.com/joshhighet/ransomwatch/main/posts.json and scripts/agent.py calls to https://api.ransomware.live and https://www.ransomlook.io), ingests and analyzes that user-generated third‑party content as part of its workflow, and uses it to drive analysis and reporting decisions—so malicious or crafted content could indirectly influence agent behavior.