Skip to main content
Back to challenges
Challenge

Google ADK Contractor Liability & Injury Analysis

Construction risk managers must evaluate complex contractual liability and safety logs to determine subcontractor indemnity obligations. Build a Google ADK multimodal agent integration with Flappy to parse legal contracts and site safety logs, classifying liability transfer with 92% accuracy.

Machine LearningHosted by Vera
Sign in free to start
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Integrate Google ADK with Gemini Vision and Flappy to extract contract indemnification clauses and worker injury reports.

Delivery guide

How work is evaluated

Evaluation

Evaluates legal indemnification clause extraction and correct assignment of liability between GC and Subcontractor.

Datasets

Challenge datasets

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

Evaluation rubric

How submissions are scored

These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.

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

liability_decision_match

Verifies decision matches court benchmark outcome

Binary check

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

Dimension 2

accuracy

Classification accuracy across test judicial decisions • target: 0.92 • range: 0-1

Binary check

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

Learning goals

What you should walk away with

  • Implement a Google ADK (Agent Development Kit) multimodal worker safety pipeline

  • Use Flappy autonomous agent framework to structure multi-step contract reasoning steps

  • Extract indemnity obligations from general contractor-subcontractor agreements

Resources and assets

Reference links and supporting material

Dataset notes

Scanned legal contracts and OSHA incident reports from construction injury litigations.

Challenge setup

Evaluates legal indemnification clause extraction and correct assignment of liability between GC and Subcontractor.

Challenge input

JSON with contract_text and injury_incident_report

Google ADK

Google's framework for multi-modal Gemini agent interactions

Flappy

Structures production agent workflows and tool calls

Evaluated output

JSON containing general_contractor_liable (bool), rationale, and clause_citation

Evaluation checks
  • Verifies decision matches court benchmark outcome
  • Classification accuracy across test judicial decisions • target: 0.92 • range: 0-1
Configured evaluation evidence
  • Benchmark: Legal-Liability-Bench
  • Accuracy target: 0.92
  • 1 public reference case
Configured runtime
  • Python execution harness
  • Python sandbox (unavailable on Versalist)

Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.

View technical recipe

Configured tools

Environment
  • Modal · Optional
Action Space
  • Google · Required
  • Google ADK · Optional
Policy Serving
  • Modal · Optional

Evaluation contract

  • liability_decision_match · Weight 1
  • accuracy · Weight 1

Metrics

  • accuracy · Target 0.92

Benchmarks

  • Legal-Liability-Bench
  • security

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON with contract_text and injury_incident_report

    Output: JSON containing general_contractor_liable (bool), rationale, and clause_citation

    Execution harness: Python

Evaluation tools

  • Google ADK · Agent SDK Framework
  • Flappy · Agent System Orchestrator

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Google ADK Contractor Liability & Injury Analysis