Skip to main content
Back to challenges
Challenge

Multi-Agent Editorial Integrity Suite with CrewAI

Create a multi-agent team using CrewAI to analyze content originality and publication standards. In light of concerns regarding copied material and restructuring in digital media, this agent team will act as a 'content integrity unit'. Agents will be assigned specific roles: a Researcher, a Fact-Checker, and a Synthesis Expert. They will use Claude Sonnet 4.6.6 to perform deep semantic comparisons and audit content workflows, ensuring that all published work maintains high craftsmanship standards.

Business OperationsHosted by Vera
Sign in free to start
Challenge brief

What you are building

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

Orchestrate a CrewAI swarm with Claude Sonnet 4.6.6 and LangWatch for editorial integrity and content evaluation.

Delivery guide

How work is evaluated

Evaluation

Evaluation of agent performance in detecting semantic similarity and authorship irregularities

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

ConsistencyCheck

Agent team converges on same audit score

Binary check

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

Dimension 2

AuditPrecision

F1 score on content evaluation • target: 0.85 • 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

  • Master CrewAI hierarchical role definition for content integrity teams

  • Integrate Claude Sonnet 4.6.6 for high-fidelity reasoning in editorial analysis

  • Deploy LangWatch for observability and drift monitoring in content audit workflows

  • Utilize Upstage SDK for advanced parsing of unstructured news article inputs

  • Implement Bito AI assistant hooks to facilitate developer and editor interaction with the agent team

  • Build a structured agent collaboration pattern using /dev/agents patterns

Resources and assets

Reference links and supporting material

Dataset notes

Dataset will be provided

Challenge setup

Evaluation of agent performance in detecting semantic similarity and authorship irregularities

Challenge input

Document raw text

CrewAI

Framework for orchestrating

LangWatch

LLM monitoring and analytics

Evaluated output

Audit report JSON

Evaluation checks
  • Agent team converges on same audit score
  • F1 score on content evaluation • target: 0.85 • range: 0-1
Configured evaluation evidence
  • AuditPrecision target: 0.85
Configured runtime
  • Docker execution harness

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

Action Space
  • CrewAI · Required
  • crewAI · Optional
  • LangWatch · Optional

Evaluation contract

  • ConsistencyCheck · Weight 1
  • AuditPrecision · Weight 1

Metrics

  • AuditPrecision · Target 0.85

Workflow tasks

  1. Task 1

    Input: Document raw text

    Output: Audit report JSON

    Execution harness: Docker

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Multi-Agent Editorial Integrity Suite with CrewAI