Skip to main content
Loading navigation
Back to challenges

Google ADK & Guardrails AI Model Hardware Standard Safe Device Router

Develop a hardware command router compliant with Model Hardware Standard (MHS) safety specifications using Google ADK and Guardrails AI. Implement physical payload validation rules that enforce strict actuator constraints, achieving zero unsafe state transitions across simulated robotic execution test suites.

EngineeringHosted by Vera
Sign in free to start

What you are building

Construct an MHS-compliant physical hardware router using Google ADK and Guardrails AI to validate and control robotic actuator actions safely.

How work is evaluated

Evaluation

Tests safety boundary enforcing for physical hardware payload validation.

Datasets

Public datasets, plus the private datasets you have access to.

How submissions are scored

Dimensions
2 scoring checks
Binary
2 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

out_of_bounds_rejection

Ensures all out-of-bounds hardware requests are explicitly rejected.

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

guardrail_overhead_ms

Validation latency per command payload. • target: 5 • range: 0-50

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

  • Implement Google Agent Development Kit (ADK) agent pipeline for MHS hardware command dispatching

  • Integrate Guardrails AI validators to enforce physical payload kinetic boundaries (velocity, torque, spatial bounds)

  • Build mock physical device driver interfaces conforming to Model Hardware Standard specifications

  • Achieve 100% detection and blocking of out-of-bounds mechanical control requests

Resources

Dataset notes

Synthetic physical motor command dataset containing 500 boundary test scenarios across robotic arm, pan-tilt head, and conveyor systems.

Challenge setup

Tests safety boundary enforcing for physical hardware payload validation.

Challenge input

JSON object containing joint angles, motor torque values, and velocity vectors.

Google ADK

Structured tool call invocation and execution framework.

Guardrails AI

Provides schema validation for physical constraints.

Evaluated output

JSON safety response with status (APPROVED|REJECTED) and risk breakdown.

Evaluation checks
  • Ensures all out-of-bounds hardware requests are explicitly rejected.
  • Validation latency per command payload. • target: 5 • range: 0-50
Configured evaluation evidence
  • Benchmark: RoboBench Safe Action Standard
  • Guardrail Overhead Ms target: 5
  • 1 public reference case
Configured runtime
  • 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

Action Space
  • Google · Required
  • Google ADK · Optional
Reward / Eval
  • Guardrails AI · Optional
Safety / Guardrails
  • Guardrails AI · Optional

Evaluation contract

  • out_of_bounds_rejection · Weight 1
  • guardrail_overhead_ms · Weight 1

Metrics

  • guardrail_overhead_ms · Target 5

Benchmarks

  • RoboBench Safe Action Standard
  • robotics

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON object containing joint angles, motor torque values, and velocity vectors.

    Output: JSON safety response with status (APPROVED|REJECTED) and risk breakdown.

    Execution harness: Python

Evaluation tools

  • Google ADK · Agent Framework
  • Guardrails AI · AI Engineering Tooling

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Google ADK & Guardrails AI Model Hardware Standard Safe Device Router