feat: add diagnostic report artifact

[yashwant-intelligaia]

Diagnostic Report Component

A structured diagnostic report component for presenting troubleshooting findings with confidence levels and actionable recommendations.

Installation

Install the component using the HAX SDK CLI:

hax init
hax add artifact diagnostic-report

This will install the component along with its required dependencies:

• @copilotkit/react-core
• zod

Usage

Basic Usage

import {
  HAXDiagnosticReport,
  DiagnosticItem,
} from "@/artifacts/diagnostic-report"

const items: DiagnosticItem[] = [
  {
    id: "1",
    suspectedCause: "Memory leak in worker process",
    confidence: 85,
    confidenceLevel: "high",
    rationale: "Memory usage increases linearly over time without recovery",
    recommendedAction: "Restart worker process",
  },
  {
    id: "2",
    suspectedCause: "Database connection pool exhausted",
    confidence: 45,
    confidenceLevel: "medium",
    rationale: "Connection wait times have increased significantly",
    recommendedAction: "Increase pool size",
  },
]

function App() {
  const handleAction = (itemId: string) => {
    console.log(`Action clicked for item: ${itemId}`)
  }

  return (
    <HAXDiagnosticReport
      title="System Health Analysis"
      items={items}
      onActionClick={handleAction}
    />
  )
}

Using Individual Components

import {
  DiagnosticReportCard,
  DiagnosticValueCard,
  DiagnosticTableHeader,
  ConfidenceValue,
} from "@/artifacts/diagnostic-report"

// Use DiagnosticReportCard for the full report
<DiagnosticReportCard
  title="API Performance Issues"
  items={items}
  onActionClick={handleAction}
/>

// Use individual components for custom layouts
<DiagnosticTableHeader
  showSuspectedCause={true}
  showConfidence={true}
  showRationale={true}
  showRecommendedAction={true}
/>

<DiagnosticValueCard
  suspectedCause="High CPU usage"
  confidence={72}
  confidenceLevel="high"
  rationale="CPU consistently above 90%"
  recommendedAction="Scale horizontally"
  onActionClick={() => console.log("Action clicked")}
/>

<ConfidenceValue value={72} level="high" />

CopilotKit Integration

Use the useDiagnosticReportAction hook to enable AI agents to create diagnostic reports:

import { useDiagnosticReportAction } from "@/artifacts/diagnostic-report"

function MyComponent() {
  const addOrUpdateArtifact = (type, data) => {
    // Handle the artifact creation/update
    console.log("Artifact:", type, data)
  }

  useDiagnosticReportAction({ addOrUpdateArtifact })

  return <div>...</div>
}

Schema

DiagnosticItem

Property	Type	Description
id	string	Unique identifier for the diagnostic item
suspectedCause	string	Description of the potential cause
confidence	number	Confidence percentage (0-100)
confidenceLevel	"high" | "medium" | "low"	Categorical confidence level
rationale	string	Explanation supporting the assessment
recommendedAction	string	Actionable next step

DiagnosticReportArtifact

{
  id: string
  type: "diagnostic-report"
  data: {
    title?: string
    items: DiagnosticItem[]
  }
}

Confidence Level Mapping

Level	Color	Range	Description
high	Green (#2ca02c)	70-100%	Strong evidence supports this cause
medium	Orange (#ff7f0e)	40-69%	Moderate evidence, needs investigation
low	Yellow (#facc15)	0-39%	Possible but requires validation

Actions

create_diagnostic_report

CopilotKit action for AI agents to create diagnostic reports.

Parameters:

Parameter	Type	Required	Description
title	string	No	Title for the diagnostic report card
itemsJson	string	Yes	JSON string of diagnostic items array

Example Action Call:

{
  "name": "create_diagnostic_report",
  "parameters": {
    "title": "Network Connectivity Analysis",
    "itemsJson": "[{\"id\":\"1\",\"suspectedCause\":\"DNS resolution failure\",\"confidence\":90,\"confidenceLevel\":\"high\",\"rationale\":\"DNS queries timing out consistently\",\"recommendedAction\":\"Check DNS server\"}]"
  }
}

Component Props

HAXDiagnosticReport / DiagnosticReportCard

Prop	Type	Default	Description
title	string	"Diagnostic Report With Actionables"	Report title
items	DiagnosticItem[]	-	Array of diagnostic items
onActionClick	(itemId: string) => void	-	Callback when action button is clicked
className	string	-	Additional CSS classes

DiagnosticValueCard

Prop	Type	Default	Description
suspectedCause	string	-	The suspected cause text
confidence	number	-	Confidence percentage (0-100)
confidenceLevel	ConfidenceLevel	-	Level: "high", "medium", or "low"
rationale	string	-	Explanation text
recommendedAction	string	-	Action button text
onActionClick	() => void	-	Action button callback

DiagnosticTableHeader

Prop	Type	Default	Description
showSuspectedCause	boolean	true	Show suspected cause column
showConfidence	boolean	true	Show confidence column
showRationale	boolean	true	Show rationale column
showRecommendedAction	boolean	true	Show recommended action column

ConfidenceValue

Prop	Type	Default	Description
value	number	-	Confidence percentage (0-100)
level	ConfidenceLevel	-	Level: "high", "medium", or "low"

Exports

// Components
export { HAXDiagnosticReport }
export { DiagnosticReportCard }
export { DiagnosticValueCard }
export { DiagnosticTableHeader }
export { ConfidenceValue }

// Hook
export { useDiagnosticReportAction }

// Types
export type { DiagnosticItem }
export type { DiagnosticReportCardProps }
export type { DiagnosticValueCardProps }
export type { DiagnosticTableHeaderProps }
export type { ConfidenceValueProps }
export type { ConfidenceLevel }
export type { DiagnosticReportArtifact }

// Zod Schema
export { DiagnosticReportArtifactZod }

Best Practices

• Order by confidence: Display items with highest confidence first to prioritize investigation
• Be specific: Keep suspected cause text concise and specific
• Provide rationale: Always include clear rationale connecting evidence to the suspected cause
• Actionable recommendations: Make recommended actions specific and actionable
• Limit items: Keep to 3-5 diagnostic items to avoid overwhelming users

When to Use

• Troubleshooting results
• Root cause analysis
• System health assessments
• Any situation requiring potential causes ranked by confidence with actions

When NOT to Use

• Simple status updates
• Single-cause issues