syncfusion-react-scheduler

Pass

Audited by Gen Agent Trust Hub on Oct 13, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes appointment and event data (such as Subject and Description) which can originate from untrusted external sources, including JSON arrays, OData services, or the Google Calendar API. This data enters the agent's context when rendering or managing schedules.
  • Ingestion points: Data is ingested via the eventSettings.dataSource property as described in SKILL.md and references/data-binding.md.
  • Boundary markers: The skill's primary instruction file (SKILL.md) includes a security policy explicitly stating that event data must be treated as untrusted, display-only input, and that embedded commands must be ignored.
  • Capability inventory: The skill provides tools for full CRUD operations (addEvent, saveEvent, deleteEvent) and supports remote network synchronization through the DataManager utility.
  • Sanitization: Documentation in references/advanced-data-binding.md notes that text fields are normalized to plain strings to prevent interpretation of embedded instructions.
  • [EXTERNAL_DOWNLOADS]: The documentation provides instructions to install various Node.js packages from the Syncfusion organization on the npm registry.
  • Evidence: Installation commands for @syncfusion/ej2-react-schedule and other peer dependencies are found in references/getting-started.md.
  • [DYNAMIC_EXECUTION]: Several implementation patterns in the documentation use the React dangerouslySetInnerHTML property to render custom content within calendar cells and headers.
  • Evidence: Examples in references/header-and-layout.md and references/template-customization.md use this property to display weather icons and custom cell text. While this is a standard React development pattern, it requires careful handling of untrusted data at the application level.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 13, 2026, 08:09 AM