Skip to main content
Loading navigation
Back to challenges

Build a Mastra AI Patent Infringement & Genomics Prior Art Workflow Agent

Multi-million dollar patent litigation in single-cell and liquid biopsy technologies demands automated prior art comparison tools. Build a TypeScript workflow agent using Mastra AI to parse patent claim language, search scientific literature prior art, and compute claim overlap scores.

Data ScienceHosted by Vera
Sign in free to start

What you are building

Develop a Mastra AI TypeScript workflow agent to compare genomic sequencing patent claims against scientific prior art.

How work is evaluated

Evaluation

Evaluates Mastra AI workflow accuracy in mapping patent claim elements to scientific prior art.

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

high_overlap_invalidation

Ensures overlap score > 0.8 correctly sets novelty_invalidated to true.

Binary check

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

Dimension 2

claim_mapping_precision

Precision of matched patent claim elements to prior art passages. • target: 0.88 • 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

  • Construct stateful workflows and agent steps using Mastra AI framework in TypeScript

  • Extract independent and dependent legal claims from genomic sequencing patent filings

  • Perform vector semantic retrieval over scientific PubMed abstracts for prior art detection

  • Calculate claim element mapping metrics and novelty conflict summaries

Resources

Dataset notes

Dataset of 20 paired single-cell/liquid biopsy patent claims and corresponding peer-reviewed research papers published prior to patent filing date.

Challenge setup

Evaluates Mastra AI workflow accuracy in mapping patent claim elements to scientific prior art.

Challenge input

JSON containing claim_text (string) and prior_art_abstract (string).

Mastra AI

TypeScript framework for modular AI workflows and state management.

Evaluated output

JSON containing overlap_score (float), mapped_elements (list), and novelty_invalidated (boolean).

Evaluation checks
  • Ensures overlap score > 0.8 correctly sets novelty_invalidated to true.
  • Precision of matched patent claim elements to prior art passages. • target: 0.88 • range: 0-1
Configured evaluation evidence
  • Benchmark: Patent Prior Art Mapping Benchmark
  • Claim Mapping Precision target: 0.88
  • 1 public reference case
Configured runtime
  • JavaScript execution harness
  • javascript 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
  • Mastra AI · Required
  • Cohere · Optional
Safety / Guardrails
  • Protect AI · Optional
Orchestration
  • Mastra AI · Required

Evaluation contract

  • high_overlap_invalidation · Weight 1
  • claim_mapping_precision · Weight 1

Metrics

  • claim_mapping_precision · Target 0.88

Benchmarks

  • Patent Prior Art Mapping Benchmark
  • biotech

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing claim_text (string) and prior_art_abstract (string).

    Output: JSON containing overlap_score (float), mapped_elements (list), and novelty_invalidated (boolean).

    Execution harness: JavaScript

Evaluation tools

  • Mastra AI · Agent Framework

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Build a Mastra AI Patent Infringement & Genomics Prior Art Workflow Agent