analyzing-web-server-logs-for-intrusion

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly instructs reading and parsing web server access logs (see SKILL.md and scripts/agent.py invoked with --log-file /var/log/nginx/access.log), which ingest untrusted, user-generated request URIs, user-agents, and referers that the agent interprets and uses to make detection/reporting decisions.