mcpbeat

Compliance Patterns

vibeeval/compliance-patterns

GDPR data handling, audit logging, data classification, retention policies, and consent management for regulatory compliance.

2k tokens
context cost
the whole folder, loaded on every use
1
files
instructions only
0
copies elsewhere
how many repositories repackaged it
521
stars on the repo
on the repository, not the skill itself

Install

one command, takes just this skill from the repository
npx skills add https://github.com/vibeeval/vibecosystem --skill compliance-patterns

The instruction itself

8 sections, as written by the author

Compliance Patterns

Data governance and regulatory compliance patterns for software systems.

Data Classification

// Tag every data field with its classification level
enum DataClass {
  PUBLIC = 'public',           // Marketing content, product info
  INTERNAL = 'internal',       // Business metrics, employee count
  CONFIDENTIAL = 'confidential', // Customer emails, order history
  RESTRICTED = 'restricted',   // Passwords, SSN, payment cards, health data
}

// Schema-level classification
interface UserRecord {
  id: string                    // INTERNAL
  email: string                 // CONFIDENTIAL (PII)
  displayName: string           // CONFIDENTIAL (PII)
  passwordHash: string          // RESTRICTED
  dateOfBirth: string           // RESTRICTED (sensitive PII)
  preferences: object           // INTERNAL
  createdAt: Date              // INTERNAL
}

// Field-level encryption for RESTRICTED data
const ENCRYPTED_FIELDS: Record<string, DataClass> = {
  'user.email': DataClass.CONFIDENTIAL,
  'user.dateOfBirth': DataClass.RESTRICTED,
  'user.ssn': DataClass.RESTRICTED,
  'payment.cardNumber': DataClass.RESTRICTED,
}

function shouldEncryptAtRest(fieldPath: string): boolean {
  const classification = ENCRYPTED_FIELDS[fieldPath]
  return classification === DataClass.RESTRICTED
}

function shouldMaskInLogs(fieldPath: string): boolean {
  const classification = ENCRYPTED_FIELDS[fieldPath]
  return classification === DataClass.CONFIDENTIAL || classification === DataClass.RESTRICTED
}

Audit Logging

interface AuditEvent {
  id: string
  timestamp: string         // ISO 8601
  actor: {
    id: string
    type: 'user' | 'system' | 'admin'
    ip?: string
  }
  action: string            // e.g., 'user.profile.updated', 'order.deleted'
  resource: {
    type: string
    id: string
  }
  changes?: {
    field: string
    oldValue: unknown       // Masked if RESTRICTED
    newValue: unknown       // Masked if RESTRICTED
  }[]
  metadata?: Record<string, unknown>
  result: 'success' | 'failure' | 'denied'
  reason?: string           // For denied/failure
}

class AuditLogger {
  constructor(private store: AuditStore) {}

  async log(event: Omit<AuditEvent, 'id' | 'timestamp'>): Promise<void> {
    const auditEvent: AuditEvent = {
      ...event,
      id: crypto.randomUUID(),
      timestamp: new Date().toISOString(),
      changes: event.changes?.map(c => ({
        ...c,
        oldValue: shouldMaskInLogs(c.field) ? '[REDACTED]' : c.oldValue,
        newValue: shouldMaskInLogs(c.field) ? '[REDACTED]' : c.newValue,
      })),
    }

    // Audit logs are append-only, immutable, tamper-evident
    await this.store.append(auditEvent)
  }
}

// Middleware: auto-audit all mutations
function auditMiddleware(auditLogger: AuditLogger) {
  return async (req: Request, res: Response, next: NextFunction) => {
    const originalJson = res.json.bind(res)

    res.json = function(body: any) {
      if (['POST', 'PUT', 'PATCH', 'DELETE'].includes(req.method)) {
        auditLogger.log({
          actor: { id: req.user?.id ?? 'anonymous', type: 'user', ip: req.ip },
          action: `${req.method.toLowerCase()}.${req.path}`,
          resource: { type: req.path.split('/')[2], id: req.params.id ?? 'N/A' },
          result: res.statusCode < 400 ? 'success' : 'failure',
        }).catch(err => console.error('Audit log failed:', err))
      }
      return originalJson(body)
    }

    next()
  }
}

GDPR Data Subject Rights

// Right to Access (Article 15): export all user data
async function exportUserData(userId: string): Promise<DataExport> {
  const user = await db.user.findUnique({ where: { id: userId } })
  const orders = await db.order.findMany({ where: { userId } })
  const activityLog = await db.activityLog.findMany({ where: { userId } })
  const consents = await db.consent.findMany({ where: { userId } })

  return {
    exportDate: new Date().toISOString(),
    subject: { id: userId, email: user?.email },
    personalData: {
      profile: sanitizeForExport(user),
      orders: orders.map(sanitizeForExport),
      activity: activityLog,
      consents,
    },
    format: 'JSON',
    version: '1.0',
  }
}

// Right to Erasure (Article 17): delete user data
async function deleteUserData(userId: string): Promise<DeletionReport> {
  const report: DeletionReport = { userId, deletedAt: new Date(), items: [] }

  // Soft-delete user profile
  await db.user.update({
    where: { id: userId },
    data: { email: `deleted_${userId}@deleted.local`, deletedAt: new Date() }
  })
  report.items.push({ type: 'user_profile', action: 'anonymized' })

  // Hard-delete activity logs
  const { count } = await db.activityLog.deleteMany({ where: { userId } })
  report.items.push({ type: 'activity_logs', action: 'deleted', count })

  // Retain orders for legal/tax compliance (anonymize PII)
  await db.order.updateMany({
    where: { userId },
    data: { customerName: '[DELETED]', customerEmail: '[DELETED]' }
  })
  report.items.push({ type: 'orders', action: 'anonymized_pii' })

  // Log the deletion itself (audit trail required)
  await auditLogger.log({
    actor: { id: userId, type: 'user' },
    action: 'user.data.erased',
    resource: { type: 'user', id: userId },
    result: 'success',
  })

  return report
}
interface ConsentRecord {
  userId: string
  purpose: string           // 'marketing', 'analytics', 'third_party_sharing'
  granted: boolean
  grantedAt: Date
  expiresAt?: Date
  source: string            // 'signup_form', 'settings_page', 'cookie_banner'
  version: string           // Consent text version (re-consent needed on change)
}

async function checkConsent(userId: string, purpose: string): Promise<boolean> {
  const consent = await db.consent.findFirst({
    where: {
      userId,
      purpose,
      granted: true,
      OR: [
        { expiresAt: null },
        { expiresAt: { gt: new Date() } },
      ],
    }
  })
  return consent !== null
}

// Guard: only process data if consent exists
async function sendMarketingEmail(userId: string): Promise<void> {
  const hasConsent = await checkConsent(userId, 'marketing')
  if (!hasConsent) {
    console.log(`Skipping marketing email for ${userId}: no consent`)
    return
  }
  await emailService.send(userId, marketingTemplate)
}

Retention Policies

// Automated data retention enforcement
const RETENTION_POLICIES: RetentionPolicy[] = [
  { dataType: 'session_logs', retentionDays: 90, action: 'delete' },
  { dataType: 'audit_logs', retentionDays: 2555, action: 'archive' },  // 7 years
  { dataType: 'user_activity', retentionDays: 365, action: 'anonymize' },
  { dataType: 'deleted_accounts', retentionDays: 30, action: 'purge' },
]

async function enforceRetention(): Promise<RetentionReport> {
  const report: RetentionReport = { executedAt: new Date(), actions: [] }

  for (const policy of RETENTION_POLICIES) {
    const cutoffDate = new Date()
    cutoffDate.setDate(cutoffDate.getDate() - policy.retentionDays)

    const count = await executeRetentionAction(policy.dataType, cutoffDate, policy.action)
    report.actions.push({
      dataType: policy.dataType,
      action: policy.action,
      recordsAffected: count,
      cutoffDate,
    })
  }

  return report
}

// Schedule: run daily via cron
// 0 3 * * * node scripts/enforce-retention.js

Checklist

  • [ ] Every data field classified (PUBLIC/INTERNAL/CONFIDENTIAL/RESTRICTED)
  • [ ] RESTRICTED data encrypted at rest and in transit
  • [ ] PII masked in all log outputs (never log raw email, SSN, card numbers)
  • [ ] Audit log: immutable, append-only, captures who/what/when/result
  • [ ] Data export endpoint (GDPR Article 15 - Right to Access)
  • [ ] Data deletion/anonymization endpoint (GDPR Article 17 - Right to Erasure)
  • [ ] Consent records with purpose, timestamp, version, and expiry
  • [ ] Retention policies enforced automatically (daily cron)
  • [ ] Data Processing Agreement (DPA) with all third-party processors

Anti-Patterns

  • Logging PII in application logs (searchable by anyone with log access)
  • Hard-deleting without audit trail (no proof of compliance)
  • Single "I agree" checkbox for all purposes (GDPR requires granular consent)
  • Retaining data indefinitely "just in case" (violates data minimization)
  • Consent version not tracked: consent invalidated when terms change
  • No data inventory: unable to locate all PII for subject access requests

How to use it

Copy the folder

Take vibeeval/compliance-patterns from the repository into ~/.claude/skills for personal use, or into .claude/skills inside a project.

Check the name does not clash

The agent identifies a skill by the name field in its header. Two skills with the same name cannot sit side by side — one of them will be ignored.