OT Cybersecurity Cyber Incident Response Agent with Google ADK
Operational technology cyber attacks can corrupt PLC firmware or force unexpected safety system shutdowns. Build an autonomous OT incident response agent using Google ADK that ingests network syslog anomalies and automatically initiates safe isolation playbooks. Achieve a 100% containment score without triggering unintended safety shutdowns on clean lines.
What you are building
Build an OT security incident response agent using Google ADK abstractions to evaluate network syslogs and trigger containment playbooks.
How work is evaluated
Evaluates incident containment speed and safety preservation during simulated cyber attacks.
Datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
safety_preservation
Ensure PLC safety system remained online during containment
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
containment_success_rate
Proportion of cyber attack scenarios successfully isolated • target: 1 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Learning goals
Configure Google ADK agents for continuous syslog stream evaluation
Design tool routines for executing firewall port isolation and PLC read-only locks
Validate zero unhandled false alarm isolations across 50 simulated baseline test scenarios
Resources
Factory network syslog logs recorded during simulated OT ransomware and unauthorized write injection attacks.
Challenge setup
Evaluates incident containment speed and safety preservation during simulated cyber attacks.
Challenge input
JSON syslog record indicating unauthorized PLC command burst
Google ADK
Offers structured tool execution loops for complex security playbooks.
Evaluated output
JSON incident response log showing isolated IP and safety status
- Ensure PLC safety system remained online during containment
- Proportion of cyber attack scenarios successfully isolated • target: 1 • range: 0-1
- Benchmark: OT Incident Response Suite (OT-IR-2025)
- Containment Success Rate target: 100%
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Shows the first task, up to three tools, and up to three evidence items.
View technical recipe
Configured tools
- Google · Required
- Google ADK · Optional
- Google Jules · Optional
- Google Jules · Optional
Evaluation contract
- safety_preservation · Weight 1
- containment_success_rate · Weight 1
Metrics
- containment_success_rate · Target 1
Benchmarks
- OT Incident Response Suite (OT-IR-2025)
- security
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON syslog record indicating unauthorized PLC command burst
Output: JSON incident response log showing isolated IP and safety status
Execution harness: Python
Evaluation tools
- Google ADK · Agent Framework
Recipe freeze marker
No recipe freeze time is recorded.