analyzing-linux-system-artifacts

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

• Third-party content exposure detected (high risk: 0.85). The required workflow mounts a forensic image and then reads outsider-authored system artifact files (e.g., user .bash_history, authorized_keys, cron/systemd files, and /var/log/*) from /mnt/evidence/... and prints their contents into the agent’s output, which would be ingested into the LLM context as analysis text.