Detection Library
lowexperimentalLinuxOCIT1485
Linux Agent Mass File Write (Seed Rule)
Baseline seed rule to detect AI agent processes writing an unusually large number of files in a short time window, which may indicate a runaway file generation loop, ransomware-like behavior, or uncontrolled data staging.
Updated Jan 10, 2025 · OCI AI Security Team
agentic-aicascading-failuremass-writedata-destructionlinuxociowasp-asi08
Problem Statement
Uncontrolled mass file creation by an AI agent can fill disk partitions, causing cascading failures across all services on the OCI instance that depend on available disk space. It may also represent data staging for exfiltration or ransomware-style encryption of existing files.
Sample Logs
{"timestamp":"2025-01-10T12:00:00Z","computer_name":"oci-worker-36","user":"agent_svc","image":"/usr/bin/python3","target_filename":"/tmp/output_00001.json","event_type":"CreateFile","note":"file 1 of 87 created in 60 seconds"}Required Fields
image
target_filename
user
computer_name
False Positives
- ·Data generation agents that legitimately produce many output files (e.g., synthetic dataset generators)
- ·Log rotation scripts that create many new log files during rotation
Tuning Guidance
Calibrate the threshold per agent type. Data processing agents may legitimately write many files. Focus alerts on writes to sensitive directories (/, /etc, /home) regardless of count.