Security Review

Procedure

Step 1 — Check @AccessControl coverage

• Every REST endpoint MUST have @AccessControl
• Verify resourceType matches the entity being accessed
• Verify actionPerformed matches the HTTP method semantics
• If skipRBAC = true is used, verify there's a comment explaining why

Step 2 — Check tenant isolation

• All TenantBase entities must have @Filter(name = "tenantFilter")
• Search for any native @Query — they bypass the filter:

  grep -rn "nativeQuery = true" openaev-model/ openaev-api/
  

• Each native query MUST include WHERE tenant_id = :tenantId or join via tenant
• Tenant relation must be @JsonIgnore — never in API output

Step 3 — Check data exposure

• New controllers must use Output DTOs — never return JPA entities directly
• Swagger annotations must be explicit for LAZY relations: @ArraySchema(schema = @Schema(type = "string")) when returning IDs
• No tenant_id in any JSON response
• No stack traces or internal error details exposed to clients

Step 4 — Check authentication & authorization

• Protected endpoints require valid session (Spring Security)
• Admin-only operations: ResourceType.UNKNOWN or explicit admin check
• Grant-managed resources: verify they're in RESOURCES_MANAGED_BY_GRANTS
• isUserHasAccess() returns meaningful logic, not just return true

Step 5 — Check secrets & credentials

grep -rn "password\|secret\|api_key\|apiKey\|token" --include="*.java" --include="*.ts" src/

• No hardcoded credentials
• Secrets in application.properties use environment variables
• No .env files committed

Step 6 — Report

Document findings using conventional comments format:

• issue (blocking): for security vulnerabilities
• suggestion (non-blocking): for improvements
• note: for informational items